':..A. FBy-Wyssiing SUBMIGROSCOPIC MORPHOLOGY OF PROTOPLASM :^ :-•:■''* I SUBMICROSCOPIC MORPHOLOGY OF PROTOPLASM SOLE DISTRIBUTORS FOR THE U.S.A. AND CANADA: ELSEVIER PRESS, INC., 402 LOVETT BOULEVARD, HOUSTON (tEXAs) FOR THE BRITISH COMMONWEALTH EXCEPT CANADA: CLEAVER-HUME PRESS, LTD., 42A SOUTH AUDLEY STREET, LONDON, W.I vy > I — ' # ■ SUBMICROSCOPIC MORPHOLOGY OF PROTOPLASM xj by A. FREY-WYSSLING PROFESSOR OF GENERAL BOTANY AT THE FEDERAL INSTITUTE OF TECHNOLOGY ZURICH (SWITZERLAND) Second English Edition No^ a i^ir ELSEVIER PUBLISHING COMPANY AMSTERDAM • HOUSTON • LONDON* NEW YORK 1953 German Edition 1938 First English Edition 1948 Second English Edition 1953 English Translation by May Hollander, Selborne ALL RIGHTS RESERVED THIS BOOK OR ANY PART THEREOF MAY NOT BE REPRODUCED IN ANY FORM (including PHOTOSTATIC OR MICROFILM FORM) WITHOUT WRITTEN PERMISSION FROM THE PUBLISHERS Library of Congress Catalog Card Number: j2-j6j9 FOREWORD ' ' LIBF?A5t:Y ^ MASS. ~ ' This monograph is the third edition of my "Submikroskopische Morphologie des Protoplasmas und seiner Derivate" published in 1938 by Gebriider Borntrager Berlin. War and post-war conditions made it impossible to republish this book in German. For that reason I was glad to accept the offer of the Elsevier Publishing Company, Amster- dam to translate the manuscript of the second edition into English. The aim of the first edition was to introduce Submicroscopic Morphology as a new branch of General Morphology. As, in 1938, the electron microscope had not yet become an instrument of biological research, that introduction was based on the results of indirect methods of investigation (macromolecular chemistry, double refraction, di- chroism, X-ray diffraction etc.), which made it possible to provide evidence of the arrangement of submicroscopic elements. In general, one indirect method alone will not produce unequivocal evidence of a structure invisible in the ordinary microscope. But a combination of several such methods made it possible to exclude certain possi- bilities. Submicroscopic Morphology, therefore, was an exciting and inspiring field of trial and error for morphologists interested in Biophysics. Since then, the electron microscope has made it feasible to photo- graph submicroscopic structures and to check the results of the indirect methods. It is a great satisfaction for the pioneers of Sub- microscopic Morphology to know that their postulates as to the struc- tures of gels, fibres etc. were right. On the other hand, our science has lost one of its attractive charms ; we no longer have the satisfacdon of inventing new methods of research and seeking the particular structural arrangement which agrees with the results given by all the available indirect methods and therefore must correspond to the real invisible structure. This romance of discovery has given place to the technical problem of obtaining objects thin enough to get the best possible image in the electron microscope. VI FOREWORD By the time the second edition appeared in 1948, Submicroscopic Morphology had become generally accepted as an important branch of the biological sciences. The morphologists who did not trust indirect methods, willingly accepted the results of electron microscopy, although electron optics are even more complicated than those of polarized light or X-rays. But the objectively visible image has always been the foundation of Morphology, and therefore research in Sub- microscopic Morphology is henceforth governed by the remarkable invention of the electron microscope. As a consequence, this third edition is centred on the results of the electron microscope; the old indirect methods, however, are treated as equally valid means of research. The polarizing microscope and even the X-ray camera are more accessible to the average biologist who is interested in iine-structures than the expensive electron micro- scope. There are several excellent monographs on electron micro- scopy, but there is no other synopsis of the value and the results of the indirect methods in Submicroscopic Morphology. In the first rush of publishing electron micrographs, many micrographs were produced which would have been discarded as mere pictures of artefacts if the conclusions of indirect methods had been considered. Where there is doubt as to the accuracy of an electron micrograph, the results estabhshed by indirect methods ought to be taken into consideration. Any discrepancies between the interpretation of the results of indirect methods and those of the electron micrograph must be cleared up before a submicroscopic structure may be regarded as definitely established. This book is written, not for specialists, but for students who are attracted to this interesting field of research. It is merely an outline and does not attempt to give full details, which should be sought in the original publications quoted. The extensive literature published on this subject since 1948 has been taken into account as far as it was possible in this condensed monograph. It shows the enormous development of Submicroscopic Morphology during this short period. Institut Fiir Allgemeine Botanik der Eidgenossischen Technischen Hochschule, Zurich. November, 1952. A. Frey-Wyssling CONTENTS INTRODUCTION: THE DOMAINS OF MORPHOLOGY i I. FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY 8 § I. Organization of Sols 8 a. Invisible Particles 8 b. Homogeneity 12 c. Concept of Phase in Colloids 15 d. Coacervation 18 § 2. Principles of Structure 24 a. Crystal Structure 26 b. Structural Chemistry 33 c. Structure of Phase Boundaries 40 d. Liquid Crystals 5^ § 3. Structure of Gels 5^ a. Chemistry of High Polymers 58 b. Structural Viscosity 64 c. Gel Structure 66 d. Micellar Theory 76 § 4. Studies in Gels 82 a. Polarization Microscopy 82 b. X-ray Analysis of Gels 96 c. Swelling of Gels 109 d. Electron Microscopy 115 e. Summary 13° n. THE FINE-STRUCTURE OF PROTOPLASM 131 § I. Cytoplasm 132. a. Molecular Constituents of the Cytoplasm 132 b. Physicochemical Behaviour of Proteins 141 c. Physical Properties of the Cytoplasm 163 d. Submicroscopic Structure of Cytoplasm 172 e. Protoplasmic Flow and Cell Polarity 186 f. Separation of the Cytoplasm into Different Phases 191 g. Morphological Principles of the Permeability Problem 197 h. Molecular Morphology of the Cytoplasm 207 § 2. Nucleus 210 a. Molecular Constituents of the Nucleus 210 b. Fine-Structure of the Nucleus 215 c. Fine-Structure of the Chromosomes 224 d. Submicroscopic Morphology of Hereditary Processes 230 § 3. Chloroplasts 243 a. Microscopic Structure of the Chloroplasts 243 b. Molecular Constituents of Chloroplasts 246 c. Submicroscopic Structure of the Chloroplasts 251 ci^^^-i VIII CONTENTS § 4. Erythrocytes 262 a. The Microscopic Structure of Erythrocytes 262 b. Molecular Constituents of the Erythrocytes 265 c. Submicroscopic Structure of Erythrocytes 266 § 5. Gametes 274 a. Spermatozoa 274 b. Eggs 276 III. FINE-STRUCTURE OF PROTOPLASMIC DEiaVATIVES 279 § I. Carbohydrates, Chitin and Cutin 279 a. Meristematic Plant Cell Walls (Cellulose) 279 b. Cutinized Cell Walls (Cutin) 293 c. The Chitin Frame (Chitin) 301 d. Starch Grains (Amylose and Amylopectin) 310 § 2. Proteins 326 a. Reserve Protein 326 b. Silk (Silk Fibroin) 331 c. Horny Substances (Keratin) 338 d. Connective Tissue (Collagen) 345 e. Muscle Fibres (Actomyosin) 352 f. Nerves (Neurokeratin and Neuronin) 360 g. Fibrillar Proteins. Recapitulation 364 RETROSPECT 371 LITERATURE 375 AUTHOR INDEX 401 SUBJECT INDEX '. 407 .^^'^rooi";.^/. i LIBRARY ? .^. t^^i.Z.^'M ^ - MASS. y.^ >•» INTRODUCTION THE DOMAINS OF MORPHOLOGY Dass ich erkeniie, was die Welt I»! Imiersten :^usammenhdlt. Goethe's Faust Morphological biology comprises the study of organs^ (anatomy in medicine, organography according to Goebel), of tissues (histology) and of cells (cytology). Together these domains form a hierarchic system, since they describe units of diminishing size in the above order. The diiferent domains defined by the concepts organ, tissue, and cell can also be characterized by the expedients which are used to make the units under investigation visible, since each of the three sciences makes use of different instruments of observation. The organographer observes with the naked eye or with the magnifying glass, the histologist with the ordinary microscope and the cytologist with the more refined immersion, phasecontrast (Zernike, 1946) or even ultraviolet microscopes. Accordingly, the range of research of organography is in general limited by the resolving power of the eye, the domain of cytology by the resolving power of the microscope (Fig. i). In biology, all that can be described with the aid of these means of observation is referred to as morpholo^. The hierarchy of morphology, however, goes beyond the resolving power of the microscope. The persistent, I might almost say the heroic, struggle with which the resolving power of the microscope has been increased (Abbe, 1879; Kohler, 1904) is the best evidence of this. Fig. I shows how the microscopic domain was widened step by step by advances in the theory and technique of optics until at about ^ In this connection, "organ" is to be understood in the morphological sense as part of an organism, and not in the physiological sense "organ = instrument", which is based on specific functions; according to that definition, single tissues, special cells or even parts of cells can also act as "organs". 2 INTRODUCTION O.I /< the absolute limit was reached for a true image of the object, due to the wavelength of ultraviolet light. Until recently, morphology was forced to remain at this limit. We have great admiration for the numerous cytologists who have worked in the limiting regions of the optical resolving power of the immersion microscope, pursuing ever finer structures, with ineffable devotion and utter disregard for their eyesight. However, if one remembers their labour and its limited prospects of success (since the actual ultrastructure of the protoplasm cannot be obtained with any certainty by microscopic means), it is remarkable how few biologists have drawn the obvious conclusion from the theory of the limit of microscopic images and have turned their attention to indirect methods of research. The resolving limit of the microscope is like the shore of a mys- terious mountain lake. On land the geo-morphologist can easily re- Fig. I THE DOMAINS OF MORPHOLOGY Organography I cm I mm Histology Cytology Micellar studies 1 Molecules Atoms I m/Li I A I lo-i 10-2 . ' IO-3 lo-"- ^ ■ « - ■ IO-* io-« ' h > ' iQ-'' io-*cm be c ■> O o d oj 9 § >s'0 "" c „ p, o ^ (/) :i 3. 6 o 0 *r\ \n. rt 2- W M D.S o '^ Q* 60 a. M g M-, o Oi (A ° a P. 2 ^2 8 E 2 g '55 J^ o 6 a O > n ctt w 2 u -macroscopic- -microscopic- submicroscopic • -amicroscopic- cognize all details of shape and size ; he can measure and photograph them. If, however, he wants to study the morphology of the bed of the lake, he will derive no benefit from a stndy of the shoreline, however carefully this may be done. Nor will it serve his purpose if he tries to THE DOMAINS OF MORPHOLOGY 3 look at the bottom through the water above it. He must not cUng stubbornly to the rocks on the shore but must free himself from the land. He must "swim", and from the surface of the water must find out indirectly with a plummet how the bed is shaped. Each fathoming provides him with a point, and the profile of the bed can then be constructed by interpolation. Until recent times, in the submicroscopic domain which lies beyond the microscopic limit, the situation was completely analogous. Views on submicroscopic morphology could be obtained only by sounding, i.e., by indirect means, and the invisible shapes and sizes could be deduced only from a combination of the various methods of research. The discovery of the electron microscope after 1938 suddenly brought the submicroscopic regions within reach. By means of electron rays the resolving power has been increased a hundredfold in one sudden leap. The surface of the water in the lake to be studied has, so to speak, been reduced to a much lower level. The precipices and gullies which had hitherto been hidden have become accessible to the in- vestigator, who is now equipped with the means whereby he can move about in this difficult province. Submicroscopic morphology has accordingly lost something of its mysterious charm. The unravelling of its secrets no longer wholly depends upon an ingenious com- bination of partial evidence obtained indirectly, as it still does in the study of the constitution of organic molecules in structural chemistry. There is now a direct means of checking the conceptions developed so far. The objective micrographs given by the electron microscope have made submicroscopic morphology very popular in biology, whereas formerly it had been left to those few biologists with a working knowledge of physics. However, the electron microscope cannot completely replace the indirect methods which have been so successful up to now. All specimens have to be dried and this may cause serious artefacts in structures like protoplasm containing 80 to 90 % water, and there exist many objects which, for technical reasons, cannot yet be imaged in the electron microscope; furthermore, irradiation by electrons represents a bombardment which, compared with irradiation by light, involves incomparably greater energies. These are apt to destroy the structures of specimens cut into insufficiently thin sections. Amicro- scopic structures, invisible in the electron microscope, may, moreover. 4 INTRODUCTION occur. The electron-optical images of biological objects should therefore be considered critically. They must be compared with the results ob- tained from the indirect methods and, in cases of contradiction, it must be made clear on which side the error lies. In this way it has been possible in some instances to show that the electron microscope had produced spurious effects. Electron microscopy should not, therefore, supersede the methods formerly appHed, such as polarization micro- scopy or X-rays analysis, but the new direct method and the valuable indirect methods must be used jointly, each acting as a check on the other, in the exploration of the submicroscopic domain. The history of this science will soon be able to celebrate its first centenary (Nageli, 1858). However, only in the last thirty years has there been enough interest to produce a continuous development of this field of research. For Ambronn, who devoted his whole life to this branch of science and who published his fundamental researches on the rod-Hke nature of the structural elements of gels in I9i6-'i7, had to carry out his work, according to his own statement, "excluded from publicity", and until his death in 1927 he considered that his was the voice of a biologist crying in the wilderness. The general lack of interest in submicroscopic problems was without doubt due to the following. Colloid chemistry had developed into a general doctrine of dispersoids. The discovery of the ultramicrbscope (Siedentopf and ZsiGMONDY, 1903) had suddenly widened the range of the sub- microscopic morphology of sols. With great enthusiasm biologists mastered the new method, but discovered with disappointment that nearly all important biological objects: cytoplasm, nuclei, plastids, cell walls, etc. are "optically empty". We know now that this is due not only to the close packing or the hydrophilic nature of the hypo- thetical particles, but also, and mainly, to the fact that we have to deal with anisodiametric structural elements, which are invisible in the ultramicroscope if only one of their dimensions is amicroscopic, even if such structural elements accumulate in loose meshworks of sub- microscopic or even microscopic dimensions. This indicates that biological gels do not at all represent disperse systems in the classical sense of colloid chemistry (see Table II). The failure of the ultra- microscope seemed to imply that these objects do not possess a sub- microscopic structure. In the meantime, structural chemistry has developed amicroscopic THE DOMAINS OF MORPHOLOGY 5 molecular morphology. X-ray analysis has provided us with exact data on the mutual position and distances of the atoms and groups of atoms in organic molecules, and this has greatly added to our knowl- edge of stereochemistry. Nowadays we know, not only the formulae of many compounds, but also, with astounding accuracy, their entire morphological structure. From the molecular region, the elucidation of the constitution of high polymers has already advanced into the submicroscopic region as a new branch of structural chemistry. In the case of polysaccharides and polypeptides, for instance, it shows that thousands of similar structural elements can be united to gigantic chain molecules which sometimes even reach microscopic lengths. Staudinger, to whom we owe this knowledge, designates this new kind of study as macro- molecular chemistry. This might lead one to believe that the link between cytological and molecular morphology has been forged and that, consequently, a special submicroscopic morphology would become superfluous. This, however, is by no means true, for, the high polymer chains can arrange themselves in more or less regular lattices which in their turn cluster together to form porous structures, interspersed with numerous capillary spaces of various sizes. Or again, they may form loose mesh- works with a totally different degree of order. Besides chains, there may occur lamellar high polymers, thus allowing for a great many possible arrangements of the submicroscopic elements. Consequently, in addition to the problems of constitution in macromolecular chemis- try, there exist morphological problems of a special kind, the de- scription of which can best be characterized as the morphological study of fine-structure. In biology this nomenclature is synonymous with the study oi micellar systems (Frey 1928b), provided the new definition on p. 81 be taken into account. In Fig. I the lower boundary line of the morphological domain of the fine-structures has been drawn arbitrarily at the limit of visibility of the smallest gold particles in the ultramicroscope. The resolving power of the electron microscope, which may yet be improved, lies for the present within the same range. This serves to show that the order of magnitude of our field of research coincides with that of classical colloid chemistry. In contrast with the isolated dispersed particles, however, the colloid dimensions do not refer to all three INTRODUCTION directions in space but, in the case of rod-shaped elements, to two dimensions only, or even to only one in the case of lamellar submicro- scopic elements, which may be clustered to form complicated systems. TABLE I MORPHOLOGY Instrument Order of Morphological hierarchy of research Scale magnitude Organs Organography Eye, magnif. glass mm scale > o. I mm Tissues Histology Microscope I Immersion ^ and ultraviolet ( microscope Micrometer Wavelengths > I /x Cells Cytology of light > O.I IX Fine-structure Micellar Electron Colloid > I m^u. studies microscope dimensions Molecule Structural X-rays Wavelength > lA structure chemistry of X-rays Atom structure Electron Electron rays Wavelength of < O.I A theory electron rays According to Table I the domain of fine structures forms a link between our present knowledge of cytological and molecular morphology. We must therefore attempt to penetrate into the study of micellar systems from these two known sides. Starting from the region of visible structures, we must resort to our knowledge of phases, while on the other, molecular, side we should apply our knowl- edge of crystal structure. Both these theories cover morphological domains which fall outside the hierarchy given in Table I. There is no upper Hmit to the dimensions of phases, although there does exist a lower hmit which we shall have to consider. Similarly, there exists no upper limit, on theoretical grounds, to the regular arrangement of atoms and molecules in crystal lattices. For this reason, we can use these abstract sciences, which are less sensitive to dimensions, as an introduction to the study of fine structures. Morphology is not an ultimate goal of science, but it represents one of its most important foundations. No physical problem can be attacked without first defining accurately the mutual positions of the various THE DOMAINS OF MORPHOLOGY 7 points in the system to be investigated. It is only after this that time can be introduced as a parameter, to pass on from static to dynamic considerations. Just so in biology. Every physiological^ research, being concerned essentially with changes in course of time, presupposes a complete knowledge of morphology. The relatio ns between the various organs and tissues can only be studied in their dependence on time if their spatial arrangement has been ascertained with accuracy. This explains the tremendous flight which the physiology of the human body has taken in connection with the development of anat omy and histology. Passing from the total organism to the elementary organism of the cell, we must expect similar relations. If, therefore, we want to study the physiology of cells successfully, we must know their morphology as thoroughly as that of the total organism. The invisible texture of the cell, however, which is the object of line structure or micellar morpho- logy, is still in its infancy. The difficulties in this field of research are great and at present we still do not know how far we shall be able to proceed. Each new gain in this direction, however, will not only augment the archives of the descriptive science of nature, but will redound to the benefit of physiology, and will in the end satisfy our thirst for knowledge. 1 Physiologji is the science of events and processes in living organisms. Both these ex- pressions clearly indicate that time is involved, i.e., they show the dynamic character of physiology. Biomorphohgy and biochemistry, on the other hand, are not concerned with time; the one describes the spatial arrangement and the other the properties of organic matter. It is only when time begins to play a part that morphology hQCome.s physiolo^ of development and biochemistry becomes physiology of metabolism which, combined, give general physio- logy, taking into account all variable quantities, i.e. space, matter and time, which are accessible to our tools of research. In view of this, we fail to see why the attribute "dynamic" is nowadays added so readily to the branches of knowledge which describe biology. A combination such as "dynamic morphology" is quite inconsistent because, by definition, morphology can do no more than describe or explain given spatial arrange- ments, whereas, as soon as changes in spatial arrangement are considered, we enter the domain of physiology. I. FUNDAAEENTALS OF SUBAETCROSCOPIC MORPHOLOGY "Le cytoplasnie propremeiit dit se present e sur le vivant comme une substance collo'idale homogene, translucide, op- tiquement vide a I'ultramicroscope. . ." GuiLLiF.RMOND, Mangf.not et Plantefol (1933, p. 386) § I. Organization of Sols a. Invisible Particles Ever since Graham (1861) showed that the pseudo-solutions which impede filtration and which nowadays we call sols contain relatively large, slowly diffusing particles, the nature of these invisible particles has been explored in all directions by colloid chemistry (Zsigmondy, 1925; OsTWALD, 1927). Demonstration and shape of the particles. Numerous methods have been worked out to distinguish and to separate the originally hypothetical submicroscopic colloid particles from the amicroscopic molecules. By means of dialysis the amicroscopic particles can be made to permeate through a semi-permeable membrane (parchment) through which the colloid particles cannot follow (Graham, 1862). This method has since been developed into ultrafiltration^ by which sols are pressed through filters with submicroscopic pores (collodion films of varying pore size) and in this way are split up into fractions of different particle sizes. Further, since most colloid pa'rticles carry an electric charge or can be charged by a change in the acidity of the sur- roundings, they can be made to migrate in an electric field to the anode or to the cathode according to their charge, and it is possible in this way to concentrate them by electrophoresis. None of these methods of indirect particle identification, however, is quite as convincing as ultramicroscopj, which makes the particles visible (SiEDENTOPF and Zsigmondy, 1903). Admittedly, the ultra- microscope does not give a true image of the colloid particles, for the I I ORGANIZATION OF SOLS 9 reason that its resolving power does not surpass that of the ordinary microscope. It merely reveals the existence of submicroscopic particles. The possibility of ultramicroscopic demonstration is based on the fact that hght incident upon small particles is scattered in all directions. In this way they become radiant (like the dust particles in a dark room where sunlight penetrates through some gap), so that the path of a beam of light in a sol is clearly traced (Tyndall scattering). The lighted sphere surrounding such a dust particle is much larger than the scattering particle itself, and an image of it can be obtained in the microscope if the distance between the colloid particles is not too small. As the objective of the microscope gives an image of planes only, optical cross-sections of the lighted spheres are imaged in the form of deflexion discs. Since the particles in the sol take part in Brownian movement, these scintillating "deflexion discs" oscillate vividly in an irregular manner. It is an impressive sight to watch these luminous spots which, in untiring movement, stand out like bright stars from the pitch-dark background. To what extent the size of the "deflexion discs" exceeds that of the particles we do not know; nor can we determine the exact shape of the particles. All the same, the ultramicroscope enables us to draw conclusions as to their circumference in cases of marked deviation from the spherical. Non-spherical particles may be oriented in a field of flow. In that case they scintillate to difterent extents according as the incident ultramicroscopic irradiation is parallel or perpendicular to the direction of flow; they show what is calJed azimuth effect. If the light falls upon the small endplane of submicroscopic rods, they scatter much less than with sideways irradiation. From such diflferences in intensity of the "deflexion discs", depending on the direction of the incident beam, the rod-hke shape of the particles can be inferred. Anisodiametric particles are usually birefringent. As they are oriented in a field of flow, sols containing such colloid particles become optically anisotropic in a velocity gradient (Freundlich, Stapelfeldt, and Zocher, 1924). Long rods are oriented at lower rates of shear than shorter ones (Signer and Gross, 1935). From measurements of the birefringence of flow, conclusions can therefore be drawn regarding the rations between length and thickness. Si^e of the particles. A clear picture of the world of submicroscopic particles can be obtained with the aid of the methods mentioned. lO FUNDAMENTALS OF SUBMI C ROSCOPIC MORPHOLOGY I But colloid chemistry was not content with these qualitative con- clusions ; it tried to obtain quantitative facts as to the size of the par- ticles. Some information was provided by ultrafiltration, but apart from that, much more accurate methods were available. If the number of particles per unit volume is determined in the ultramicroscope, the particle size can be calculated from the con- centration of the sol. Moreover, there exist mathematical relations between Brownian movement (Einstein's formula), velocity of sedi- mentation (Stokes' formula) or diffusion on the one hand, and particle size on the other. These make it possible to determine the diameter of spherical colloid particles. The ultramicroscope plays an important part in these investigations (Zsigmondy, 1925), since the particles have to be observed when counting or measuring the Brownian movement. In many cases, however, the colloid particles cannot be observed ultramicroscopically, not only because their dimensions are frequently too small but, above all, because their refractive power is often only sUghtly different from that of the dispersing medium, so that light scattering is insufficient. This usually applies to biological sols with their organic colloid particles, which means that the limit of visibility of these sols in the ultramicroscope is reached long before that of inorganic sols (compare Fig. i, p. 2). The method of sedimentation is free from this difficulty, because the change in concentration of the solution as a result of sedimentation of the particles can be determined by analytic means or, still more simply, by the change in refractive index. Moreover, the sedimentation velocity can be increased at will by applying stronger centrifugal forces. The ultracentrifuge, which was developed by Svedberg (1938a) into an instrument of the highest accuracy and great power (cen- trifugal fields which are 750,000 times that of the gravitational field!) allows of the determination of particle weights down to amicroscopic molecules. The various methods referred to have revealed much of the morphology (size and shape) of submicroscopic particles, so that the electron microscope has only confirmed by direct micrographs the results obtained by indirect means. Fig. 2 represents a series of submicroscopic particles of biological importance, facilitating comparison with the microscopic and amicro- scopic regions. The size and shape of the particles were determined by ORGANIZATION OF SOLS II Fig. 2 PARTICLE SIZES (PARTLY FROM STANLEY, 1938a, b) measuring scale Mol. weight 0 in I . Erythrocyte .... z. Bacterium coli . . . 3 . Bacterium prodigiosum 4. Treponema pallidum 5 . Small-pox virus fj 6. Chicken plague virus . 7. Megatherium bacteriophagus . 8 .Yellow fever virus 100 m/x 9. Gene, calcul. accord, to Muller(i935) 10. Tobacco mosaic virus. . . . 1 1. Foot and mouth virus. . . . 12. Glycogen, according to HusEMANN and Ruska (1940) 1 3 . Haemocyanin from Octopus . 14. Smallest ultramicroscopically visible gold particles, accord- ingtoZsiGMONDY(i925) . 15. Horse haemoglobin . . . . 16. Ovalbumin (Svedberg 1930) 17. Saccharose 18. Hydrogen molecule . . . . 2300- 10° 300- lO" 23- 10° 4.3-10" 33-io« 43- 10" 0.4- 10® 1.5- 10^ 2.8- 10° 2.7- 10° 69- 10* 5-5 40- 10' 4-3 342 0.5 2 0.2 7500 3000 750 200 Length {mfi) 175 90 38 22 20 12.3 10 10 6000 1000 18000 125 430 64 i.o 12 FUNDAMENTALS OF SURMIC ROSCOPIC MORPHOLOGY I the methods mentioned and in many cases also by the electron micro- scope. It is seen that there is a continuous transition from the lifeless amicroscopic molecules to the living cells at the limit of microscopic visibility. The smallest particles which exhibit phenomena of life (self- multiplication) are in the submicroscopic region. Theoretical biology^ being concerned with the definition and the essence of life, is therefore called upon to give serious attention to our branch of morphology. On the other hand, these colloid particles often give the impression of consisting of uniform, chemically well-defined substances, and the biochemist attributes molecular weights to them which, depending on the size of the particles, may assume fantastically large values. b. Homogeneity Real solutions containing amicroscopic particles are designated as uniform or homogeneous from a physico-chemical point of view. Sols, however, are not considered as uniform; they are heterogeneous. The concept of homogeneity applied here is essentially different from the optical homogeneity which plays such an important part in microscopy. A medium is optically homogeneous when its constituent parts have the same refractive index, so that it is impossible to establish their boundary line by means of light. Physico-chemical homogeneity, however, requires that two parts taken from the object shall be identical, not only in their behaviour towards light, but also in all other properties. This will be the case if the particles are similarly arranged throughout the whole object (Figs. 3-7). Several homogeneous arrangements of particles are possible. The structural elements can be arranged irregularly, like the molecules of a liquid or gas. The distances between the particles are not all equal, but if we proceed through the mass along a straight line, the average distance found will be constant, and equal volume elements will on the average contain an equal number of particles. Such arrangements are called statistically homogeneous in contrast to the distribution of the atoms in a crystal, which are arranged in a certain pattern. As all distances in a given direction are identical, this is called a lattice arrangement. The spacings can be equal in three directions which are mutually perpendicular; in that case the lattice arrangement is isotropic (Fig. 4). Or else, the spacings are different in different di- ORGANIZATION OF SOLS 13 f . •. . • • • .\ • • • • / • • ^^* •/• • • • 9 ^, • • • . • • • . • • • « • • . • • • . . • ^ ^ • • • e ^ • \ • • f* • \» . • • • •/ • • 1 • • • • - • • • • • • • ."- • •/ • • • • • • • • • • • • • • • ••••• • c • • • • •/ " • ^ ^« • • • / • • ^ • • • I* • •! • • • V • • / • • • • • • • • o • • • • • • • '/"'-^^ • • • •/ • • • « • • -V _•__ ^ • ••••• • • • •! • •••••••••• • ••••/*•••*. • • • •.• ^ • • • f^M • • • ^ • •••!••• •••'••••9' , I ^_,/ I 1 ^ • • •(• •••••••• •!• • I I I I Fig. 3 Fig. 4 Fig- 5 Fis. 6 Fig- 7 Homogeneous arrangements Fig. 3. Statistically homogeneous distribution - Fig. 4. Homogeneous isotropic lattice - Fig. 5. Homogeneous anisotropic lattice - Fig. 6. Statistically homogeneous distribution of polar particles - Fig. 7. Homogeneous lattice arrangement of polar particles. rections, in which case the lattice arrangement is anisotropic (Fig. 5). The homogeneous lattice arrangement has in common with the statistically homogeneous arrangement that equal volumes contain an equal number of particles. With anisotropic arrangements it is not sufficient to compare volumes of equal size; they must also have the same orientation. For, if from Fig. 5 instead of circles we draw two congruent rectangles with different orientations, the properties of one of these rectangles will be different from those of the other on account of the different distribution ot lattice points with respect to the length of the rectangle (the linear thermal expansion of the long side of the two rectangles, for instance, will be different). The necessity of taking orientation into account becomes particularly apparent if 14 FUNDAMENTALS OF SUBM I C RO SC OPI C MORPHOLOGY I Fig. 8 %^®^ at •^ 2 ^ a: ^ cA Fig. 9 Fig. lo Fig. II Fig. 12 Fig. 13 Homogeneous states of the compound ABj. QA, 0 B, ^ E. Fig. 8. Solid - Fig. 9. Liquid - Fig. 10. Gaseous - Fig. 11. Homogeneous solution of EBj in AB2. Mixed crystals (A, E) Bg. Fig. 12. Homogeneous - Fig. 13. Heterogeneous. polar particles such as, for instance, water molecules are arranged homogeneously. Fig. 6 shows such particles in a statistically homoge- neous distribution and Fig. 7 gives an example of an arrangement in a lattice which has identical spacings in three directions. From these considerations we derive the following definition of homogeneity: an object is homogeneous if equal and equally oriented parts, taken arbitrarily from the object, possess the same internal structure. This implies that all the parts thus compared have the same physical and chemical properties. An important condition in these considerations is the order of magnitude of the volumes to be compared. Physico-chemical homo- I ORGANIZATION OF SOLS I5 o-eneity requires that they shall be oi submicroscopic dh/iens'ions .Y\xQ.vc\X&t- nal structure, therefore, refers to the arrangement ot atoms, ions and molecules, which in Figs. 3-7 have been indicated by points or arrows. It follows from this definition that sols cannot be homogeneous. For, if in a sol we consider submicroscopic volumes of sufficiently small size, the one may contain a colloid particle, while the other may merely contain the solvent, i.e., the dispersing medium. In contrast to sols, not only are all pure substances homogeneous, whether in the solid, liquid or gaseous state (Figs. 8-10), but so also are real solutions, provided the solute consists of amicroscopic particles (Fig. 11). If, however, differences in the concentration, for instance concentration gradients, occur in the solution, it is heterogeneous. Similarly, either homogeneous or heterogeneous mixed crystals can originate from a solution or melt, according as the two components can unite to a crystal lattice in a regular or in an irregular distribution (Figs, iz and 15). Colloid solution having been recognized as heterogeneous, the further question arises whether the colloid particles themselves may be considered as homogeneous. To answer this question we must deal shortly with the phase theory, which treats of relations between homogeneous states. c. Concept of Phase in Colloids According to the thermodynamical definition, any homogeneous state is called a phase. Figs. 8-12 thus picture the structure of phases. Fig. 13 representing not a homogeneous phase but a heterogeneous system of AB2 and EBg. The colloid particles were formerly believed to be homogeneous and the dispersed particles were therefore designated as dispersed phase and the surrounding liquid as dispersing medium (Fig. 14). Thus a sol represents a two-phase system. The study of the structure of colloids need not, of course, be confined to the liquid state. Dispersions of liquid or solid particles in liquid or solid media (emulsions, suspen- sions, etc.) are known in the microscopic domain. We may also expect to find them in the submicroscopic world. Since, however, the par- ticles in such dispersions are no longer visible, colloid systems were designated as dispersoids. In this way an attempt was made to charac- terize, not only the organization of sols, but in the most general sense l6 FUNDAMENTALS OF SUBMICROSCOPI C MORPHOLOGY I that of all colloids, as will be clear from the following system (Wo. OsTWALD, 1909). Systematics of dispersoids. According to the theory of dispersions, each of the three states of matter, solid, liquid or gas, can occur either as dispersing medium or as dispersed particles (Fig. 14), so that 3^ = 9 combinations are possible (Table II). Fig. 14 shows how in these systems the dispersed part I is distributed in the dispersing phase 11. TABLE II DISPERSOID SYSTEMS, ACCORDING TO WO. OSTWALD, 1909 Dispersing medium Dispersed portion Dispersoids Solid Solid Grain-structure Solid Liquid Drop-structure Solid Gas Bubble-structure Liquid Solid Suspensoids Liquid Liquid Emulsoids Liquid Gas Foams Gas Solid Smoke Gas Liquid Mist Gas Gas On the strenght of the definition of phases it was originally believed that the dispersed part I was homogeneous. In the dispersoids, how- ever, this leads to difficulties. Often it was doubtful whether a dispersed phase was liquid or soHd. For, suppose the dispersoid particles become smaller and smaller until they contain only a few molecules, then it would be difficult to decide whether they are solid or liquid. Liquid drops may be taken to be homogeneous, whereas it is very difficult to prove this of solid suspended particles. It was only by the introduction of X-ray iiiethods in colloid chemistry that the particles of certain dispersoids, for instance gold and silver sols, could be proved to possess a crystal lattice and, therefore, to be really homogeneous. With increasing degree of dispersion, however, the homogeneity of a crystal lattice also becomes questionable. For, the energy ot the points lying at its surface is different from that of the points inside the lattice, because they are no longer surrounded on all sides by equivalent fields of force (Fig. i6). In the case of liquids this sives rise to surface tension. For instance, in the smallest gold particles ORGANIZATION OF SOLS 17 which can be measured by X-ray methods (Scherrer, 1920), 200 of the 380 Au-atoms, i.e., more than half the total number, lie at the sur- face of the crystals. With decreasing particle size, of course, an even higher percentage of atoms lies at the surface, until, with 14 or still less, all the atoms lie at the surface (face-centred cube. Fig. 25, p .27). Thus one can no longer speak of a homogeneous phase in the case of atoms Phase I Phase E Fig. 14 Fig. 15 Fig. 16 Fig. 14. Colloid chemical concept of phase. I Dispersed phase (colloid portion), 11 dispersing medium. The inhomogeneity of boundaries reigns throughout the system - Fig. 15. Thermodynamical concept of phase. The homogeneity of the phases reigns throughout the system. - Fig. i6. Inbomogeneous surroundings of the lattice points of boundary' planes (face-centered cubic lattice). that are not similarly surrounded on all sides. It is only by a still further increase in dispersion that finally a homogeneous, molecularly dispersed solution of Au-ions is obtained. On the other hand, it has been ascertained (Zsigmondy, 1925, p. 39) that the homogeneous primary particles of suspensoids can cluster together to form bigger heterogeneous secondary particles (compare Fig. 73, p. 104) without any fundamental change in the properties of such sols. This strengthened the opinion that the properties of sols and other colloids were not decided by the inner structure of the particles. Since with increasing dispersion the surface of the particles increases considerably in proportion to their mass, colloid chemistry has developed much more into a science of surfaces. The properties and reactions of colloids have been elucidated to a great extent by the study of surface reactions. Whereas the phase theory is concerned with the equilibrium between different phases and is able to predict under what conditions phases cease to exist (dissolution) or new phases appear l8 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I (separation into two layers), classical colloid chemistry is interested in the first place in phase boundaries (capillary chemistry according to Freundlich, 1922). Thermodynamics require that all parts of a phase have exactly the same energy content. This is only realized, however, when the phases are so extended that the irregular distribution of energy at their surface, i.e., the inhomogeneity in the immediate neighbourhood of the phase boundary (Fig. 16) can be neglected (Fig. 15). Thus, the classical phase theory has to forego all considerations concerning phase boundaries (compare Figs. 3-11, 13, 14) because of their inhomoge- neity, and its laws only apply to homogeneous regions of at least micro- scopic dimensions. The properties of colloids, on the contrary, are determined in the first place by the inhomogeneity of the phase boun- daries, the predominant effect of which is due to the very large surface. For this reason it has been suggested by Ostwald (1938) that the definition "dispersed phase" should be avoided, and that we should speak of the "colloid portion" of the dispersoid. The phase theory once seemed to hold out promise of explaining the formation of new phases (separation into two strata, formation of vacuoles) or the disappearance of phases (melting-in) in biological systems. From the above, however, it is clear that the phase theory does not hold good in colloid chemistry, since it has been developed by emphasizing the homogeneity of the phase and neglecting the specific properties of surfaces, while conversely, in cytological systems, homogeneity usually fails and the surfaces are of quite outstanding importance. Bungenberg de Jong and his fellow-workers have elucidated the principles according to which visible boundary layers can appear and disappear in those heterogeneous systems to which the phase theory does not apply. In his theory of coacervation Bungen- berg DE Jong has summarized the rules which govern these phe- nomena. d. Coacervation In the separation of a sol into two non-miscible parts, the dispersing medium and the dispersed portion often do not separate completely. Flakes are formed which still contain a certain amount of dispersing medium and therefore remain suspended. For this reason the floccu- lation is usually reversible. If, however, such flakes collect into small ORGANIZATION OF SOLS 19 drops or into a coherent liquid layer, we have to do with a phenomenon, for which Bungenberg de Jong (1932) introduced the term coacervation (Fig. 20); in English: piling up (acervus — pile). Hydration. The colloid particles in a sol are solvated, which means that molecules of the dispersing medium adhere to the particle. In the special case of water, this solvation is designated as hydration, since in that case water molecules are bound by the colloid particle. The W H \ /- / 1 \ x" ^^ Fig. 17 Fig. 18 Fig. 19 Fig. 17. Model of a water molecule and scheme of dipoles - Fig. i8. Hy- dration of an isoelectric colloid particle - Fig. 19. Hydration of a charged colloid particle (from Pallmann, 193 i). attraction is brought about by electrostatic forces, for, in a water molecule the electric charges are not distributed uniformly, because the two positive hydrogen atoms are separated in space from the doubly charged negative oxygen. For that reason a water molecule in an electric field behaves like a molecular rod with two different electric poles and is therefore designated as a dipole (Fig. 17). Similarly, in a colloid particle the electric charges are usually not distributed uni- formly, not even if the particles are isoelectric, i.e., if their positive and negative charges cancel each other so that outwardly they appear neutral. In Fig. 18 a particle has been sketched, the negative charges of which are situated towards its surface. This has a polarizing effect on the water molecules in the immediate neighbourhood of the particle. These water molecules follow the particle in its Brownian movement as the so-called solvation or hydration layer.' If the colloid particle is not neutral but carries an excess negative or positive charge as a result of dissociation of H- or OH -ions, the swarm of oriented 20 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I dipoles surrounding the particle will be correspondingly larger. This is why the hydration of colloids reaches its minimal value at the isoelectric point. The binding forces which attract the water dipoles decrease with increasing distance. Thus the swarm of water molecules which are hampered in their free movement becomes less dense in the outer layers until in the end one reaches without noticeable transition the region of the freely moving dipoles of the dispersing medium. In the solvation layer the density of the water therefore decreases ex- ponentially, in much the same way as the density of the atmosphere with increasing distance from the earth. As thert is no sudden transition from the hydration layer to the free water, such hydrophilic colloids are very stable. The particles show no tendency to cluster together; in a way they "have no surface at all", their surface energy is zero (Fig. 2oa). Dehydration. If water is withdrawn from the diffuse solvation layer, the difference between bound and freely moving dipoles becomes noticeable. The water layer around the particle now acquires a surface (Fig. 2ob) and if two such dehydrated particles meet, the surface . ■ No surface • z'. -■--.V . ./•.;. ." '■ . "^ " '• ''.y ■^•^T"'-*. * "\ • 1 ■•■■% sXM*** ' 1 ^ \--:> '^."'•■'. V ' \- ■ • . ■*/' .* */ . " •>^- * . - . Surface Flocculation .Coacervafe droplet Coacervafe layer Fig. 20. Dehydration of colloid particles (from Bungenberg de Jong, 1932). a) Diffuse hydration layer, b) definite hydration layer, r) incipient coacervation. energy which tends towards a minimum value will cause the surround- ing water layers to unite. The colloid particles, however, cannot come into direct contact with each other because of their solvation layers. But they no longer possess separate layers, for these have all united into a single liquid sphere. If the number of particles united in this way becomes so large that they form a microscopically visible conglomeration, one speaks of flakes or flocculates. These can further cluster into drops (microcoacervation) and finally into a liquid layer (macrocoacervation). Thus coacervates are liquids rich in colloid which have been separated by means of dehydration. I ORGANIZATION OF SOLS 21 In the coacervate the distribution of colloid particles is statistically uniform, as in the original sol, although their concentration has been increased. If the colloid particles are considered as dispersed phase, their state has not been changed in the coacervation process; and yet clearly a new phase boundary is formed between a layer rich in colloid and one poor in colloid. This example shows how vague is the con- cept of phase in colloid chemistry. For that reason the hydrophilic sols and the coacervates originating from them are sometimes called quasi-homogeneous phases, since the distribution of the particles is completely uniform and the particle size is liable to decrease to mole- cular dimensions. The dehydration of colloid particles illustrated in Fig. 20 can be brought about in various ways; for instance, a rise in temperature, which accentuates the contrast between bound and freely moving water, will often suffice. Usually, however, use is made of dehydrating substances such as sahs (salting out) or aliphatic alcohols or acetone. Such substances, which disturb the stability of the sol and increase the tendency to separate, are called sensitizers. Besides salts and organic liquids, colloid solutions may also be used as sensitizers if they com- pete with the particles of the original sol to bind the free water and thus cause dehydration. The dispersing medium which is separated from the coacervate is called the equilibrium liquid (Fig. 21), for, following changes in tempera- ture or composition in the system, water is taken up or given oflF by the coacervate. The situation is, therefore, analogous to the separation in a mixture of phenol and water (p. 46). Coacervates can be regarded as a solution of water in the colloid (swelling) and the equilibrium liquid must then represent a solution of a small amount of colloid in water. In the example given in Fig. 21, however, the gelatin is insoluble in alcohol-water and the concentration of the colloid in the equilibrium liquid is practically zero. Here the analogy therefore ceases since, in- stead of a reciprocal solubility, there exists only a one-sided adsorption of water by the colloid. The reason why gelatin is completely in- soluble below its melting point will be made clear on p. 73. Coacervates of homopolar substances have been studied by Mme Dobry (1938, 1940). Discbarge. In biological systems the colloid particles are seldom neutral; usually they are electrically charged. Particles carrying opposite 22 FUNDAMENTALS OF SUBMIC ROSC OPI C MORPHOLOGY charges tend to unite, but because of their solvation layers can only- approach each other to a certain extent. The attraction is counteracted by the hydration as by a spring (Fig. 22) and thus no coagulation takes place which would annul the charges, but again a coacervate occurs which now contains particles of opposite charge. So, in addi- Equilibrium liquid Coacervate Fig. 21. Coacervation of gelatin at 41"^ C. Isoelectric gelatin sol + alcohol as sensi- tizer. Equilibrium liquid = solution of water and alcohol. Coacervate = gelatin + small amounts of water + alcohol. v^AA\AAA Fig. 22. State of stability of colloid particles (from BuNGENBERG DE JoNG and Bonner, 1935). Attraction by opposite electrical charges (arrows). Repulsion by solvation layer (spring). tion to sensitizers, electric charges are apt to cause coacervation. For this the sols must have opposite charges; e.g., gelatin (positive) and gum arable (negative) or lecithin (positive) and nucleic acid (negative). In this case the aggregation is designated as complex coacervation, since two oppositely charged kinds of particles take part in the flocculation. In many cases colloid particles can be made to reverse their charge by adding neutral salts, when the familiar valency rules apply, viz., on the addition of polyvalent cations, negative particles change their sign more easily according as the valency of the cation is higher, while positive particles behave in a similar way with respect to polyvalent anions. Negatively charged phosphatides, for example, reverse their charge on the addition of CaClg. In the sol, the phosphatide particles which have already become positive and those which have so far re- mained negative attract each other, and in this way a separation occurs which has been called autocomplex coacervation, because in this instance similar but oppositely charged particles attract each other. Morphologically the coacervation shows many features which have their counterpart in the phenomena occurring in cells. In the first place the vacuolization calls for mention. If, in a system consisting of equili- ORGANIZATION OF SOLS 23 brium liquid and suspended coacervate droplets, the equilibrium is modified as a result of changes in temperature or composition in the direction of a further dehydration (heating, addition of more sensiti2er), vacuoles appear in the droplets. These represent separated equilibrium liquid which has remained inside the coacervate droplets (Fig. 23). Fig. 23. Vacuolization by lowering the temperature of coacervate drops consisting of gelatin sol + resorcin (from Bungenberg de Jong, 1932). Probably vacuolization by dehydration is comparable with the forma- tion of vacuoles in the cell, since, in that case too, liquid is being separated from the plasma colloids. Apart from this striking analogy, Bungenberg de Jong (1932) mentions other models for cytological differentiation on the basis of observations with coacervates. When mixing sols of gelatin, gum arable and nucleic acid from yeast, two complex coacervates arise, in addition to equihbrium hquid, one of which consists mainly of gelatin and gum arable, the other being composed chiefly of gelatin and nucleic 24 FUNDAMENTALS OF S UBMI C RO S C OPIC MORPHOLOGY I acid. Their partition is such that the first always contains the second in the form of enclosed droplets. This can easily be demonstrated, as the negative nucleic acid coacer- vate can be selectively stained by alkaline dyes such as methyl green. This is regarded by Bungenberg de Jong as a model for a nucleus imbedded in cytoplasm. Personally, however, I do not believe that such comparisons are admissible, since both nucleus and cytoplasm possess a structure, whereas the liquid coacervate droplets are com- pletely amorphous. For that reason, the picture suggesting the resem- blance to the cell may be incidental and should therefore not be used in analogy to cytological phenomena. There would otherwise be too great a temptation to over-simplify the relationships between cyto- plasm and nucleus. The nuclear changes in karyokinesis, for instance, cannot possibly be attributed to changes in hydration or electric charges alone. These phenomena are attended with complicated struc- tural alterations. Whereas the early upholders of the theory of coacervation were principally concerned with the surfaces of the colloid particles with their solvation layers and electric charges, attempting to gain more knowledge of the structure of boundary layers (see p. 40 and 267), their studies were later extended to include the inner structure of coacervate systems. In biological objects we have to assume that the coacervate has a submicroscopic gel structure (Bank, i 941). Therefore, apart from a knowledge of boundary structure, we are also in need of deeper insight into the inner structure of colloid particles and coacer- vate flocculates. In order to advance in this direction we must appeal to structural principles. § 2. Principles of Structure By structural principles we mean the laws governing the mutual posi- tions of atoms, ions, and molecules. The positions of the atoms in the molecule are studied by structural chemistrj, which in this respect appears as a morphological science. For example, when we represent the carbon atom by its 4 valencies or a benzene ring by the well- known hexagon (Fig. 24), these are morphological illustrations based on certain properties of these substances. The exact location of the valency bonds in space and the distances between the atoms remained 2 PRINCIPLES OF STRUCTURE ZJ unknown for a long time, and there was a certain arbitrariness in the use of valency lines as regards their direction and length (cf. Fig. 35 b, p. 38). Today, however, the data needed for an exact morphological representation are known, and, if written in a suitable way, at least the simpler chemical formulae actually do represent molecular models, which have been projected on to a plane. We owe our knowledge of the exact distances and h x^ directions chiefly to X-ray analysis. X-rays ^ j. ^ ii' m enable us to measure dimensions of the order | hc^;;^^ joh of magnitude of their wavelength (e.g., copper ^ radiation: X = 1.54 A), if identical distances Fig. 24 are often repeated and act as a lattice, causing interferences which can be photographed and thus made macroscopically visible. It is, therefore, the principle oi repetition^hxch. has opened the door to the morphology ofmolecular structure. The more regularly the given distances are arranged, the more accurately can the absolute values and directions be determined. From the considerations relating to ho- mogeneity it follows, therefore, that in gases, liquids (Fig. 9-11, p. 14) and solutions the morphology of the molecules cannot be determined by means of X-rays, though an exception to this rule is provided by solutions of very large molecules which in their own construction show a certain periodicity (for example carbon chains). In such cases, however, the measurements are often ambiguous, because the mole- cules are not orientated in fixed directions. The most reliable values of atomic distances, often attaining almost incredible precision (up to i7oo of I '^)» have therefore been determined in crystal lattices. For a quantitative determination of the arrangement of the atoms in a mol- ecule one must necessarily make use of phases which possess a structure. Amorphous phases without structure, such as liquids and real solutions, are not suitable for the elucidation of such morpholo- gical relations. In this respect, biological conditions are highly unfavourable. Al- though the protoplasm must be presumed to have a structure, it is not governed by the principle of repetition with sufficient consistency to permit of X-ray analysis. Granted that periodicity plays an im- portant part in all living matter as regards time and, to some extent, also spatial arrangement; yet a strictly periodic order presupposes an equilibrium of forces and this is opposed to life, which depends on iG FUNDAMENTALS OF SUBMI C RO S COPI C MORPHOLOGY I movement and the maintenance of non-equilibria. As soon, however, as chemical substances are withdrawn from the metabolic processes, the ordering forces can intervene and form periodic structures, as, for instance, with the skeleton substances cellulose, chitin, collagen, keratin, etc. Therefore, to study the structure of protoplasm, other methods should be applied which, however, are partly based on the results of the investigations on crystal structure. For this reason this important branch of morphology must be briefly touched upon. a. Crystal Structure Lattice. The essential nature of lattices is determined by the fact that certain locations of points, which in the more simple cases are identical with the centre of gravity of the atoms, periodically repeat themselves in three given directions in space. These directions coincide with the axes of the crystallographic system. The distance from one point to the next identical one is designated as the identity period or spacing. Depending on the crystallographic system, the spacings are the same in either three (cubic) or only two directions (tetragonal, hexagonal, rhombohedral), or they are different in all three dimensions (rhombic, monoclinic, triclinic). The reg- ularly repeated points form an array of points. Displacing such a row by constant amounts in a direction either perpendicular or obliquely to its own direction, we obtain the lattice plane, while finally the crystal lattice results from displacing such a plane. If a point in the lattice is moved in the three principal directions, each time covering the identity period involved, and if the three vectors obtained are completed to a three-dimensional parallel- epiped, we obtain the so-called elementary or unit cell of the crystal lattice. In analogy to a gas molecule, which represents the smallest unit with all the chemical properties of the gaseous phase, the unit cell is the smallest unit which still shows all physical and symmetry properties of the crystal. It may contain one or several molecules (and in the case of high polymers even parts of molecules). We are, therefore, dealing with a geometrical concept and by no means with a chemical one. If the unit cell is decomposed into its elements, the crystalline properties are lost. As the base cell possesses all the properties of the crystal, and this crystal can be obtained by displacing the elementary unit in the principal directions, structure analysis aims at de- termining the dimensions and the symmetry of the base. Its shape is de- termined by three identity periods a:b:c in Angstrom units, to which in monoclinic and triclinic systems one must add the angle ^, or the angles a, /3, y formed by the edges of the unit cell. The macroscopically determined proportions between the axes of the crystals agree with the proportions between the dimensions of the unit cell, provided analogous planes are considered. X-ray analysis measures the distances between the lattice planes. In the PRINCIPLES OF STRUCTURE 27 case of crystals showing a high degree of symmetry (cubic system), the lattice points are identical with the points of intersection of symmetry planes and their distances can therefore be calculated from the distances in the X-ray diagram. In the case of lattices having a lower degree of symmetry, however, the situation of the points in the lattice planes is not determined unambiguously by symmetry elements; they possess certain degrees of freedom. Accordingly, the determination of the structure with the aid of the distances in the X-ray diagram, alone, is not possible; additional mea- surements of the intensity of the interferences are then required. In this case, however, the position of all lattice points in the unit cell can often be only approximately determined. (Niggli, 1929, 1941/42). Fig. 25 Fig. 26 Crystal lattices. The encircled points belong to the unit cell Fig. 25. Gold, a = 4.07 A, # Au - Fig. 26. Sodium chloride, a = 5.60 A, # Na, O CI Figs. 25 and 26 represent two of the best-known lattices, viz. that of the element gold and of the compound sodium chloride. Both lattices are cubic: this means that the dimensions and shape of the unit cell are determined by a single identity period a which is the same in three mutually perpendicular directions. Once the spacing a has been determined by means of X-rays, the volume a^ and, from the known density of the crystalline substance, the weight of the unit cell can be calculated. Dividing this weight by the absolute weight of the atom or molecule in question ( = atomic or, as the case may be, molecular weight/LoscHMiDT's number 0.606 • 10-^), one finds the number of atoms or molecules in the unit cell. For example, the elementary cell of gold contains 4 Au atoms, that of sodium chloride 4 Na- and 4 Cl-ions. These points have been encircled in Figs. 25 and 26; the other points marked on the planes of the cube are to be considered as having originated from the encircled ones by a simple translation, thus belonging to a neighbouring unit cell. The lattice type of gold is termed face-centred because the points of intersection of the diagonals of the faces are all occupied by atoms. Numerous elements, such as Ag, Cu, Al, Pb, etc., crystallize in accordance with the same scheme, though with different identity periods. In the NaCl type of lattice, which is 28 FUNDAMENTALS OF SUBMICROSCOPIC MORPh6lOGY I found in several binary compounds (NaF, KCl, PbS, etc.) with different values of a, two of such face-centred cubic lattices overlap. If the atoms of a crystal lattice are not represented by distinct points, but by spheres touching each other, their space requirement related to the volume of the unit cell can be calculated. It is then found that of all possible crystal lattices the cubic face-centred lattice of Fig. 25 has the closest possible packing. The volume of the spheres amounts to 0.74 of the total space available. There is another possibiHty of closest packing where the arrange- ment of the spheres is hexagonal (hexagonal space-centred lattice). The ratio of the axis is a:c = 1.63:1 and the space required exactly the same as in the cubic closest packing (0.74). In other types of close packing the space requirement is always smaller than 0.74. For instance, in the space- centred cubic lattice the spheres fill only 0.68 of the volume of the unit cell. Primary valency lattice. Next to the geometrical relations between the points in the crystal lattices, the forces which keep the atoms together are of primary importance. The purely geom.etrical consideration of the lattice is quite independent of this. As soon, however, as one is interested in the reason why certain distances in a lattice are great and others small, this question must be considered. In fact, the lattice forces are of a varied nature. Actually, in the examples given, the forces are different. In Fig. 25 similar atoms, in Fig. 26 oppositely charged ions attract each other. In both cases primary valencies act as lattice forces which can join together uncharged as well as oppositely charged particles. In the first case one speaks of a homopolar lattice, in the second of a heteropolar or ion lattice. The morphological similarity of these two types of lattice is due to the fact that in both cases the construction of the lattice is founded on the rules of the theory of co-ordination. According to Werner's chemistry of complexes, each atom is surrounded by a fixed number of neighbouring particles, either 4, 6, 8 or 1 2, depending on volume conditions. This theory, based originally on the composition of salts containing crystal water [e.g., Ca(H20)gCl2] and other complex salts, has also proved useful in the elucidation of crystal structures of other compounds and of the elements. In fact, in Fig. 25 each Au-atom at the corners of the cube is surrounded by 1 2 neighbouring atoms and in Fig. 26 each Na-ion by 6 Cl-ions or, vice versa, each Cl-ion by 6 Na- ions. The theory of co-ordination has led to another fundamental re- cognition which has become of the greatest importance to the sub- PRINCIPLES OF STRUCTURE 29 microscopic morphology of organic compounds. It has been shown that the lattice points in Figs. 25 and 26 represent only the centres of gravity of the atoms. The range of their electron orbits, however, extends over such large volumes, that these can be represented by spheres touching each other in the lattice (Figs. 27-29). A crystal Fig. 27 Fig. 28 Fig. 29 Co-ordination numbers (fromMAGNUS, 1922) Fig. 27. Number 12; e.g., Au (Au),2 in crystallized gold - Fig. 28. Number 6; e.g., Na(Cl)e in sodium chloride; Fe(CN)6 as ion - Fig. 29. Number 4; e.g., CCI4, C(C)4 in diamond. lattice, therefore, which is kept together by main valencies is much more closely packed than the common pictures suggest. Unfortunately, the representations in space obtained by drawing continuous spheres ■ instead of lattice points are not very illuminating, whereas in a plane this procedure can be applied with great success (comp. Fig. 3 1, p. 34)- The atomic distances in the lattices of elements correspond, therefore, to the atomic diameters and in binary compounds they represent the sum of the radii of the two partners (Goldschmidt). In this way it has been possible to determine the volume occupied by various atoms and at the same time to find an explanation for the different co-ordina- tion numbers. E.g., four Cl-atoms combined in a tetrahedron to- gether enclose a space which just corresponds to the size of a silicon atom; this accounts for the co-ordination number 4 in the compound SiCl4. Of the smaller fluorine atoms, however, we need 6 spheres to obtain the space occupied by one Si-atom. Hence the co-ordination number 6 (SiFg). If the lattice contains homopolar valency bonds, the distances be- tween the atoms, or the diameters of their spheres, show a surprising 30 SUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I constancy, not only in simple compounds, but also in very complicated ones. In the heteropolar ion lattices a disturbing effect occurs because of the opposite charges of the two partners. The ions have a polarizing effect upon each other, which may lead to deformations of the electron orbits (Fajans, 1923, 1925) in those cases where the symmetry of the lattice does not exclude such effects, as, for instance, in the lattice of NaCl (Fig. 26). The ions can then no longer be represented by spheres; they represent dipoles similar to the water molecules in Fig. 17 (p. 19). The result is that ion lattices often possess Httle symmetry and that the atomic distances between given partners are subject to certain fluctuations, depending on the circumstances. Fortunately this does not apply to the molecular structures of organic compounds which always have a homopolar character; the distances found in certain compounds can therefore be transferred with perfect confidence to other ones, so that one can speak of distance rules. In Table III a number of atomic distances are given as determined in organic crystals by means of X-rays. In these considera- tions the hydrogen atoms must be neglected, as they do not scatter X-rays ; nor do they seem to have a perceptible influence on the dis- tances between the atoms. Table III, for example, shows that in single bonds the atom radius of carbon, r,;, amounts to 0.77 A and that of nitrogen, r^^, to 0.71 A. In spite of the larger atomic weight of nitrogen, its sphere of action is smaller than that of carbon. It is also seen that the sphere of influence of the carbon atoms is decreased by double bonds. Each valency in an organic molecule corresponds to a definite amount of energy (Meyer and Mark, 1930). In the combustion of the homologous paraffins, for instance, the heat of combustion per mole increases by a definite amount for each new C-atom introduced; this value amounts to about 70 kcal. The energy equivalents for the other compounds mentioned in Table III have been determined in a similar way. It will be apparent that with decreasing distance between the C-atoms the energy content of the different bonds increases. To sum up, it can be said that in the main valency bonds which play a part in the structure of protoplasm, distances of 1-1.5 A and bond energies of the order of 100-200 kcal occur. Molecule lattice. In addition to homopolar main valency lattices and heteropolar ion lattices we must consider molecule lattices. If the PRINCIPLES OF STRUCTURE TABLE III DISTANCES AND MAIN VALENCY FORCES BETWEEN THE ATOMS IN ORGANIC COMPOUNDS 31 Distance in A Energy-equivalent Crystal lattice Bond accord, to kcal (Meyer- Stuart, 1954 Mark, 1930) Diamond .... Aliphatic C— C 1.54 71 Graphite Aromatic C^^C 1.42-1.4S 96 Stilbene Double C=C 1-35 125 Ca-carbide .... Triple C^C 1.19 166 Carbonic acid . . . Ketone C=:0 1.05-1.15 203 Polyoxymethylene . Oxygen bridge C — O 1.49 Urea ; hexamethyl- ene tetramine . . Amino C — N 1. 53-1.48 valency of an atom species corresponds to the co-ordination number with regard to another atom species (as, e.g., in CH4), the mutual saturation of the valencies excludes the possibility of unlimited lattices such as those shown in Figs. 25 and 26 (p. 27). Although such molecules no longer possess free valencies, they can still be arranged in a crystal lattice (see Mark and Schossberger, 1937). The binding forces, however, are now of a different nature; in contrast to the pri- mary valencies they are called secondary valencies. They are explained in theoretical physics by means of dipole moments, in much the same way as the orientation and attraction of water molecules by an ion (see Fig. 19, p. 19). In practice these forces between the molecules cause the cohesion. The secondary valence forces are, therefore, identical with the Van der Waals cohesive forces. In molecule lattices they are of the same nature as in liquids and they can therefore be derived from the heat of sublimation or vaporization of the compound. It then be- comes apparent that each atom or radical occurring in the structural formulae of organic chemistrv contributes a certain amount to the cohesion. At a first approximation the cohesion of a molecule species is composed additively of these partial contributions, and can be calculated by adding up the various increments, in exactly the same way as the molecular volume (according to Kopp's rule), the molecular 32 FUNDAMENTALS OF SUBMIC ROSCOPI C MORPHOLOGY TABLE IV COHESIVE FORCES BETWEEN ORGANIC GROUPS, ACCORDING TO MEYER AND MARK I93O Groups Molar cohesion kcal/mole Aliphatic C: methyl and ^ methylene groups S Ether bridge Amino group Carbonyl group Aldehyde group Hydroxyl group Carboxyl group — CHg and = CHj CH2 , — CH 0 NHo CO CHO OH COOH 1.78 0.99 1.63 5-53 4.27 4.70 7-25 8.97 weight or the molecular refraction. Accordingly, the contribution of the characteristic groups to the cohesion has been denoted as mo/ar cohesion (Meyer and Mark, 1930). For example, the heat of vaporiza- tion of ethyl alcohol, which amounts to 10 kcal per mole, is additively composed of the molecular cohesions of CH3, CHg, and OH. The values concerned can be found in Table IV. This table shows that, in neighbouring molecules, methyl and methylene groups and also oxygen bridges attract each other only slightly. The attraction between amino and ketone groups is twice as large and, in the polar hydroxyl and carboxyl groups, the cohesion assumes quite considerable values. None the less, all the values for molar cohesion are 10 to 100 times smaller than the energy equivalents of the main valency bonds, and accordingly the secondary valency bonds are at least 10 times weaker. Consequently, whenever secondary valencies play a decisive role in the crystal lattice, the distances are much greater than those between atoms bound by primary valencies. In organic crystals, therefore, in which both bond types occur : primary valencies inside the molecule (intramolecular) and secondary valencies between the molecules (intermolecular), the lattice distances are essen- tially of two different orders of magnitude. PRINCIPLES OF STRUCTURE 33 b. Structural Chemistry After the discovery of stereoisomery^ structural chemistry learnt to distinguish between different positions of the substituents to the carbon atom. At first, the results of this interesting science (Werner, 1904; Freudenberg, 1933) were little more than qualitative and re- ferred mainly to the directions radiating from the C-atom. Quantitative determinations of distances along these directions were not yet possible. The results of crystal structure, however, determine not only qualita- tively but also quantitatively the relative positions of the atoms in space. The starting point for the new development in structural chemistry was the crystal lattice of diamond, which crystallizes in the cubic Fig. 30. Diamond lattice, a) Unit cell, h) projection. system. Its unit cell is a cube containing 8 C-atoms, 4 of which belong to a face-centred cube as in the case of gold, while the four remaining atoms are situated on the body diagonals halfway between the corners of the cube and its centre (Fig. 50a). Thus the unit cell contains, as it were, 4 central atoms surrounded by 4 neighbouring atoms at the corners of a tetrahedron, in conformity with their co-ordination num- ber (Fig. 29). If this three-dimensional lattice is projected on to its base. Fig. 50b is obtained, which shows the arrangement of valency lines commonly used in organic chemistry ! Thus the usual scheme of the quadrivalent carbon (Fig. 24, p. 25) is morphologically correct if it is considered as the projection of a tetrahedron. According to X-ray analysis, the lattice period of diamond, i.e., the edge of the cube, measures 3.55 A. It follows that the distance between the lattice points on the face diagonal is 1-3.55 -Vz = 2.51 A; the shortest distance between two C-atoms on the body diagonal is 5-3.55 -^3 = 1.54 A. It is in this simple way that the C — C-distance corresponding to the sphere of action of a C-atom in an aliphatic bond has been calculated (Table III, p. 31). 34 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I If a plane is drawn through two body diagonals, the arrangement of lattice points obtained is as represented in Fig. 31a. In this cross- section the C-atoms are joined by a zig-zag line whose links enclose the so-called tetrahedron angle of 109°. 5. On parallel planes, further ar- rays of such zig-zag chains are found, one of which has been represented by dotted lines. It is linked up with the other two by primary valencies. .^ , I / 7.45^ 'CH, J^ /^ V r'&f TH. Y Tcwp \ ' y \ V ^ Fig. 31. «) Diamond lattice (primary valency lattice) as compared with Fig. 30a by 45° inclined, b) Paraffin lattice (molecule lattice); V = valency angle = 109°. 5 Aliphatic compounds (chain lattice). The zig-zag arrangement described is fundamental to the morphology of saturated carbon compounds; for it has been found that all aliphatic molecules represent such kinked chains. In paraffin molecules, for instance, the increase in chain length for each additional C-atom is 1.27 A instead of 1.54 A. It can easily be calculated that this is in conformity with the zig-zag chains show- ing the tetrahedron angle. In this way two carbon atoms reach a spacing of 2.54 A, which is the intramolecular period of the paraffins (Hengstenberg, 1928; MiJLLER, 1929; Halle, 193 i). In Fig. 31b it is shown how, by parallel alignment, such chains combine into the rhombic crystal lattice of the paraffins. It seems PRINCIPLES OF STRUCTURE 55 paradoxical that the soft, plastic paraffin crystals should have a lattice structure so similar to the diamond model represented by Fig. 31a. Notwithstanding the apparent analogy, however, there exist funda- mental differences which explain the differences in the physical be- haviour of the two substances. In particular, the lattice of the paraffin crystals is built much more loosely. This is caused by the fact that cleavage< plane homogeneous lattice region <> ^l <> I S 2 a b Fig. 32. Aliphatic chains, a) Molecule lattice; b) chain lattice. these crystals possess not a main valency lattice, but a molecule lattice. The chains are joined by Van der Waals forces only, since the CH2 groups are able to bind only two neighbouring groups by primary valencies. Thus in the paraffin lattice we have two types of distances : molecular distances of the order of magnitude 5 A and atomic ones of the order of magnitude 1.5 A (Fig. 31b). The fact that in the diamond lattice all C-atoms touch each other explains its great density and hardness. The paraffin lattice, on the other hand, has a much lower density and layers of molecules can be shifted with respect to each other with relative ease (Fig. 32). This accounts for the soft- ness and plasticity of paraffin crystals. As long as the paraffin chains are short, they easily crystallize into a molecular lattice. This leads to crystals in the form of flakes, which 2,6 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I can be cleft along the base (Fig. 32a). When, however, the chains grow to great length, it becomes increasingly difficult to arrange the terminal groups in fixed planes, and crystallization takes place as pictured in Fig. 32b. Here no rigorous lattice order prevails, since a lengthways displacement of one chain with respect to another, over distances equal to some intramolecular spacings, i.e., only a fraction of the chain length, does not affect the lattice structure. This is because, owing to their multiplicity, the smaller spacings inside the molecule (2.54 A in the case of paraffins) overshadow the periodicity of the end groups. These arrangements of long chains are called c/yam lattice. It is significant that the chains cannot revolve around their longitudinal axis ; if they could, there would be no lattice order. The cross-section of the chain lattice is, therefore, homogeneous, but inhomogeneities, which are indicated in Fig. 32b by the end groups, occur lengthwise, leaving only small homogeneous lattice regions. CH, .CH. I \ 3CH2 CH2 CH2 H2C4 H2C4 H2/ CH2 CHj o) \ 5CH2 V H2 .H, 'CHj H3C6 b) Fig- 33 Fig- 34 Fig. 33. Graphite lattice - Fig. 34. Hexane. a) Conventional structural formula; b) morphologically correct formula; c) ring constellation, supplement to the valency angle v = 70°. 5. Aromatic compounds {layer lattice). Unlike the aliphatic compounds, the aromatic ones cannot be derived from the structure of diamond. Their structure is similar to that oi graphite. This modification of carbon crystallizes in the hexagonal system and possesses a crystal lattice as represented in Fig. 33. The carbon atoms form rings containing 6 atoms, which are linked together in an uninterrupted plane. Thus at each lattice point 3 primary valencies are engaged. The fourth valency is distributed among the neigh- bouring atoms as in the benzene ring (Fig. 24, p. 25). Accordingly, as a result of the larger bond energy, the C — C-distance is reduced to 1.45 A (see Fig. 33). As all primary valencies are thus engaged in a plane, the 2 PRINCIPLES OF STRUCTURE 37 resulting main valency layers are united into a lattice by weaker secondary valencies. The distance between the layers (3.41 A) is therefore considerably larger than that in the rings. A structure in which the lattice forces and spacings within a plane are so different from those in a direction perpen- dicular (or nearly perpendicular) to this plane is called a layer lattice. Com- pounds of this lattice type always crystallize in the form of flakes and are as a rule easily split along the base (mica, serisite). Many benzene derivatives and other aromatic compounds (naphthalene, anthracene, etc.) belong to this class. The division into aliphatic and aromatic substances is therefore not only based upon their chemical behaviour, but it also has a morphological background in that the one tends to crystallize into a chain lattice, while the other shows a strong tendency towards the development of a layer lattice. Cyclic compounds. The structural formulae of aliphatic chemistry are found to be very similar to molecular models if the valency angle between two successive C-C bonds are taken into account. A chain such as hexane should therefore be kinked instead of straight (Fig. 34a, and b). Molecules which do not form part of a crystal lattice, but can freely move about in the gaseous or dissolved state, are subject to the so-called free rotation of the groups around the direction of the valency lines. In Fig. 34a rotation would not give rise to a new structure. In kinked chains, however, the free rotation means that, for instance, group I in Fig. 34b need not necessarily lie in the plane of drawing with 2 and 5 ; it can be located anywhere on the perimeter of a cone which has its apex in group 2 and whose apical angle is the supplement of the valency angle. Among these possibilities there is one special case in which groups 4 and 5 are turned through 1 80°, thus resulting in a ring-shaped model. It is not difficult to see that this can easily lead to cyclic compounds. Fig. 34c shows why rings of 5 or 6 atoms are formed preferentially: the supplement (70°. 5) of the valency angle is contained somewhat less than 6 and somewhat more than 5 times in 360° (5 -70°. 5 = 3 5 2°. 5; 6 -70°. 5 = 423°). The different forms which a molecule can assume are called its constellations; so Figs. 34b and c represent two different constellations of the same molecule hexane. Other atoms besides carbon can also occur in the ring (heterocyclic rings). Let us here briefly discuss the example of sugar, which is so important in biology. The monosaccharides, which formerly were considered as "open" chains (Fig. 35a), have been shown to contain a heterocyclic ring with an oxygen bridge. In glucose this is usually a T-5 bond, often represented in the manner of Fig. 35 b. The formula, 38 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I however, is not true to reality, since the C-O distance in it is unduly large. Haworth (1925), therefore, writes sugar as an equilateral hexagon or pentagon, according as to whether the oxygen bridge is situated between the carbon atoms 1-5 (derivatives of pyranose) or CH=0 H> \ : t \ Water - 20 40 60 80 100% — ^Phenol Fig. 42. Diagram of miscibility of the water/phenol system (from Rothmund, 1898). Abscissa: from left to right content of phenol in % of weight. Ordinate: Temperature in °C. PRINCIPLES OF STRUCTURE 47 of hydrophobic (i.e., water-repelling) groups in the molecule. On the other hand, the latter groups determine the solubility in homopolar liquids (solvents for fatty substances), such as hydrocarbons, carbon tetrachloride, carbon disulphide, ether, chloroform, benzene, olive oil, etc., and for this reason are designated as lipophilic groups. Table VI gives a survey of the various hydrophilic and lipophilic groups TABLE VI HYDROPHILIC AND HYDROPHOBIC GROUPS Hydrophilic (lipophobic) dipole character (often tendency to form ions) Lipophilic (hydrophobic) homopolar S ^\OH carboxyl % —OH hydroxyl ^ — C\tt aldehyde "2 C — O carbonyl ^ — NHj amino S NH imino 1 C^NH, ^^^^« ^ — C<(qj^ imido • — SH sulphydryl ^ 00 Oh C •■—* 'o bJO G decomposed ,, (supercr.) -> decomposed, infusible rising temperature The ease with which a mesophase is changed into an isotropic liquid is a function of the chain length. This is apparent, in particular, from Vorlander's researches (1936). With increasing chain length it becomes increasingly difficult to attain the amorphous liquid state, because finally the melting point II is such a high temperature that the chain molecules are decomposed before the mesophase is converted into a real liquid. With still greater chain lengths the substance does not fuse at all, because the molecules are subject to degradation be- fore becoming movable. In this case, therefore, the cohesive forces be- tween the very long chains are stronger than the main valency bonds in the chain molecule: the molecular structure breaks down before the lattice disintegrates. Substances which cannot be changed into the amorphous state, because the ////^r-molecular forces in the lattice or in sheaves (mesophase) are larger than the /«/r<2-molecular binding forces, are called super -crystalline (Vorlander). A survey is given in Table VII; the substances (2) and (3) occur as mesophases at certain intervals of temperature. Fig. 45 shows a series of molecules of increasing chain length which 54 FUNDAMENTALS OF SU BM I C RO S C OPI C MORPHOLOGY I corresponds to the general plan of Table VII. The striking fact in this series is that the addition of only a single pair of members to the chain results in such radical changes in the physical properties. It is to be noted that this holds good only for para-substituents in the benzene ring, leading to one-dimensional chain molecules. i) p-Azoxybenzene 36° cryst. -solid Z^ am. liq. o 2) p-Azoxyphenetolc 134" il'5 crvst. -solid 5^ cryst. liq. ^ am. liq. ^-/ II ^— O 5) p-Azoxy-azobenzene 226'' crvst. -solid *^ cryst. liq. ^ dccomp. -N=N-< >-N = N^ >-N = N 6 4) p-Azoxy-disazobenzene cryst.-solid -^ decomp. o Fig. 45. Series of chain molecules which aggregate to mesophases (cf. Table VII). Myelin for //IS. Cytologists are more famihar with the birefringent semi-liquid tubes, designated as myelin forms because they were ob- served for the first time with myelinated nerves (Fig. 179, p. 362). When water is added to such nerve fibres, adventitious threads issue from their sheath. They bend and curl and finally grow into irregular entanglements. The active substance causing these structures is the lecithin in the myelin sheath, for exactly the same phenomena are observed when water is added to isolated lecithin, especially if this is liable to decompose. Although the myelin forms are particularly strik- ing in organic phosphoric acid compounds, similar tubes emerge PRINCIPLES OF STRl^CTURK 55 from the alkali salts of oleic acid when these are wetted. Very beautiful myelin forms were obtained by Gicklhorn (1932a) in the cell sap of the well-known AlUuni epidermal cells by adding ammonia or sodium hydroxide (Fig. 46). The variety of shapes in these peculiar structures is beautifully demonstrated in Nageotte's microphotograph atlas (1936, No. 434). The myelin forms are usually designated as liquid crystals. It Fig. 46. Myelin forms in the epidermal cells oi Allium (from Gicklhorn, 1932a). should be pointed out, however, that there is a fundamental difference between these structures and the crystalline liquid state mentioned above. For, in the latter we have to deal with a special aggregate state of a uniform substance, i.e., a system consisting of one com- ponent only, whereas at least two components take part in the formation of myelin forms. In the examples mentioned, one of these components is water. It is further essential that the molecules, which here again must have a chain-like structure, be not homopolar as in Fig. 45, but heteropolar, i.e., they must contain a hydrophilic and a lipophilic pole. The hydrophilic group in oleic acid is the carboxyl group, that in lecithin is the choline. If the conditions mentioned are realized, myelin forms may occur, provided the molecules are sufficiently mobile. The apparent growth is due to water absorption; it is, therefore, a matter oi swelling: the hydrophilic groups are surrounded by water, while the hydrophobic groups are drawn away from the surface. The resulting orientation in the case of lecithin is represented in Fig. 47a; the lecithin underlying this scheme is a /5-lecithin (see Fig. 93, p. 138) in which the phosphoric acid is attached to the OH-group in the middle of the glycerol molecule. Obviously, the water penetrating into the 56 FUNDAMENTALS OF SUBMI C ROSCOPI C MORPHOLOGY I lecithin causes the molecules to arrange themselves in layers which are similar to surface films, except that there are no mono- or oligo- molecular layers but huge, microscopically visible structures consisting of bimolecular lamellae. If the length of the pair of overlapping lecithin molecules is about 50 A (Trillat, 1925/27), a wall of a myelin tube 5 (i in thickness consists of some 1000 double layers (Fig. 47b). Water continues to be absorbed until all the hydrophilic groups are saturated. T o o o „ 000 o .''■■'.'.'.'■','< „ r.\'.',','.',v o ° °- ° t > to o o r ■■■■■■■' ■! O O .■.'.■.'.■■■.■.■■ I o _ o 00 L O o o o o o !f^?^„° "o^'^tC c O O g "o O o o o o o o o o o ° O n ^ O 10 o o o o " B ----.1 ^Q O O ^ O op B.' ■'■'-.''■^.'■.'-'i V r.^.-. ■■-■■■.■■ . B J o o o o o o L o o o o o 00 Q o o (o o o o o o ■■■.'.v,'. I o O o o Lecithin °°°H,0°°° t c.50^ Fig. 47. Myelin forms of lecitiain a) Submicroscopic structure. Hatched, hydrophilic; black, lipophilic part of the fork- shaped lecithin molecule, b) Microscopic image and optics. thus causing further growth of the tubes. In course of time the myelin forms therefore traverse the whole field of view under the cover glass of the microscopic preparation. It can be proved by optical means that the lecithin molecules in the myelin tubes are perpendicular to the surface. For, in a flowing solu- tion (see p. 90) the lecithin molecules appear to be optically positive. The myelin tubes, however, are optically negative with respect to their long axis. From this it follows that the lecithin chains must be oriented perpendicular to the tube axis. Bear and Schmitt (1956) mention a formula (p. 86) from which the double refraction n^ — n^ of the cylindrical myelin tube with its optical axis in radial direction can be computed. For the myelin forms of lecithin in Ringer solution the authors fijid n^ — n^ = 0.0039 (Schmitt and Bear, 1937). On further absorption of water the lamellar structure of the myelin forms becomes PRINCIPLES OF STRUCTURE 57 63.5^^-- increasingly pronounced. Finally, the positive intrinsic double refrac- tion of the molecules is overcompensated by the negative double refraction due to the lamellar texture (see p. 87,) and the sign of the myelin birefringence is reversed (Nageotte, 1936). The absorption of water can be followed by means of X-rays. The dry myehn substances obtained from evapor- ated benzene solu- tions give X-ray in- terferences which cor- respond to twice the chain length (lecithin and cephalin 44A, ster- ol 34 A, sphingomyelin and cerebroside 65- 67 A; ScHMiTT and Palmer, i 940) . If water is added to these lipids, the X-ray periods are enlarged and so allow of an evaluation of the thickness of the water lamellae formed. It can be seen from Fig. 48 that the original period of 63.5 A of mixed nerve lipids has be- come 150 A at a water content of 75%. This implies a water layer of 86 A between the bimolecular lipid layers. The myelin forms offer a good example of the manner in which complicated microscopic structures can result from a simple arrange- ment of submicroscopic entities. They show, however, that no co- ordinated growth is possible as a result of such a process, for the myelin forms "grow" at random aimlessly in the substrate and the final outcome is a chaos rather than an illustration of organized life (Fig. 46). Fig. 48. Water intercalation between bimolecular lipidic films. Size of the adsorbed water layer with increasing water content. The black points correspond to o%, 25%, 50%, 6-j°o and 75% of water content (from Schmitt and Palmer, 1940). 58 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I § 3. Structure of Gels a. Chemistry of High Polymers Polymerisation and condensation. In about 1920 Staudinger drew attention to the fact that in the high-polymer natural substances the struc- tural units which can be obtained from them by hydrolysis are interlinked by primary valency bonds (Kekule bonds). He first proved the correctness Monomer H H Polymer chain H H H H c=o -C— O— C— O— C— O— C— O— C— o H Formaldehyde (oxymethylene) CH = CH., H H H H H Polyoxymethylene _CH— CH,— CH— CH.,— CH— CH,— CH— CH,— CH— Styrene / Polystyrene CH2=C— CH-CH„ — CH,— C = CH— CH„— CH,— C = CH— CH,— CH^— CH3 Isoprene CH3 CH3 Caoutchouc (polyprene) Fig. 49. Polvmerization of this point of view in synthetic products. Fig. 49 shows some of his polymerizations. It is seen that the monomer molecules always contain double bonds, one of which interacts with another molecule and thus links two monomer molecules together. If this process is repeated, long chain molecules are formed whose growth would be theoretically unlimited if the possibility of further addition did not diminish with increasing chain length and the sensitivity to oxygen (and the like) of the giant molecules formed did not become considerably enhanced. For the present, however, the factors limiting the chain length will not be considered, and the polymer chains will simply be denoted by "open" formulae. Polymerization pro- cesses are particularly successful if the monomer contains a system of conjugated double bonds, as e.g., in isoprene, i.e., if double bonds alternate with single bonds. The terminal double bonds may then give rise to inter- linking with those of neighbouring monomer molecules, while the central single bond is converted into a double bond. In this manner unsaturated 3 STRUCTURE OF GELS 59 high-polymer compounds are formed, such as rubber in the case con- sidered here. Apart from this type of chain formation, high molecular weight sub- stances may also be formed by etherification of alcohohc groups (Fig. 50) or by a process of esterification between carboxylic and hydroxylic groups with elimination of water. This way of interlinking is distinguished as condensation from the polymerization of unsaturated compounds. It leads to equally long molecules; the chains are then, however, no longer all-carbon chains like those in polystyrene or rubber, but always contain oxygen atoms as interconnecting links. When polyvalent alcohols react with each other, no chain-like, but net-hke or even spatial giant molecules are formed, such as probably occur in the insoluble huminic acids and in the insoluble cutins (see p. 295). By way of introduction, however, we shall confine the discussion to the somewhat simpler conditions in the high-polymer carbo- hydrates with linear chain molecules. The high-polymer molecules may become so large as to assume the properties of colloid particles. Staudinger (1936a) designates these giant molecules as macromolecules and the branch of science dealing with their constitution and chemical behaviour as niacromokcidar chemistry. Polysaccharides. The same principles by which disaccharides are formed (see Fig. 35/37, p. 39), govern the formation of polysaccharides, which are of outstanding importance in plant physiology. Here too, the monoses are interlinked by 1-4 oxygen bridges with elimination of water, and this polycondensation may embrace a large number of monomer molecules. In cellulose the successive links of /3-glucose are rotated through 180°. In starch, however, the a-glucose residues can interact without being rotated (Fig. 50). The cellulose chains have a digonal screw axis as an element of symmetry, contrary to the starch chains, which have not. Consequently, the cellulose molecules are more stable and straightened out, whereas the starch molecules tend to become more convolute because they are less symmetrical. This morphological difference is doubtless one of the reasons for the difference in behaviour between starch and cellulose. Possibly it is also responsible for the tendency of the starch molecule towards branching (see Fig. 1 5 2b, p. 3 1 1). The mannans occurring in corozo nut and in the rhizomes of Amorphophalhis konjak (see Fig. 160) can be derived in a similar way from mannose as starch and cellulose from glu- cose. The two monoses differ only in the different position of the H- and OH-groups at the second C-atom. For the chain in mannan from corozo nut, locahzed in the cell waU, Meyer and Mark (1930, p. 168) 6o FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY assume /5-glucosidic bonds, while it seems likely that to Konjak mannan, being a reserve substance, a starch-like structure with a- glucosidic bonds should be assigned. It is highly significant that the pectic substances, which are held to be responsible for the coherence of plant tissues and which contain poly- OH > O^OH CH20H ^-Glucose OH OHy 1 OH CH2OH cX-G/ucose OH. OH \VHOhN CH^OH C(-Monri06e OH oih-o"^ COOR CH2OH ,^ty CH2OH 0— CHpOH CH2OH '^:x°- CH2OH CH2OH Starch CH2OH -0- CH20H CH2OH CH2OH CH2OH c(-Mannan C(_-Galaciuronic acid (R^CHs or H) COOR o—^^o^ \ — A-o COOR COOR Polygalocturonic acid ^^^' O— Fig. 50. Polysaccharides galacturonic acid as a basic material, also have the structural principle of polysaccharides. Here the -CHgOH side chain of the monose ring is replaced by the carboxyl group -COOH. The pectins are therefore capable of salt formation. Polygalacturonic acid is soluble in water, but its Ca-salt is not, so that this polyacid can be precipitated by Ca ions. Part of the carboxyl groups is esterified with methanol (Deuel, 1943). It is interesting to note that the methylation does not interfere with the solubility in water, because methyl groups bound to oxonium oxygen obtain an induced polarity so that they partly lose their lipophilic character and become hydrophilic. The monomer of the pectic acid is a-galacturonic acid. As in a- galactose, the hydroxyl groups of the first and fourth C-atoms are not situated on the same side of the pyranose ring (Fig. 50); the a- glucosidic linkage causes a rotation of succeeding chain members. 3 STRUCTURE OF GELS 6l In crystalline sodium pectate the screw axis is not twofold as in cellu- lose, but threefold (Palmer and Hartzog, 1945). The crystallizing tendency of pectic substances is much smaller than that of cellulose; in the plant it occurs in the amorphous state only (Wuhrmann and PiLNiK, 1945). The pentosans, which come partly within the hemicellulose class, have a similar structure to that of the polysaccharides already de- scribed, except for the absence of the side chains, i.e., the sixth C- atom. If in cellulose or polygalacturonic acid this group is replaced by H, we obtain the xylan chain or a polyarahinan. The polysaccharides demonstrate strikingly how slight morpholog- ical variations of one and the same structural principle may give rise to substances which behave quite differently from a physiological point of view. Chain length of high polymers. According to Staudinger, all high polymer chains terminate in end groups. Unfortunately, so far the terminal groups of none of the high molecular weight natural substances are known; the chains are therefore preferably written in "open" formulae (Fig. 50). Contrarily, in comparatively short synthetic chains the end groups, hence the molecular weights, of the products can be determined. If foreign atoms, such as, for instance, iodine form the terminal groups, such determinations can be easily performed. If, however, the chains are terminated by OH- groups, the accuracy of this so-called end-group method diminishes rapidly with increasing chain length. In polyoxymethylene dimethyl ether this method can be successfully applied up to a degree of polymerization of about 100. The methods of freezing point depression and rise of boiUng point, commonly used in molecular weight determinations in substances of low molecular weight, cannot be applied to high polymers, as the effects are too small. On the other hand, the molecular weight, and thus the chain length of high polymers, can be measured by osmotic means, in which case it must be taken into account that Van 't Hoff's law does not apply rigorously to molecules of so great a volume. Corrections similar to Van der Waals' b-correction in the equation of state of gases must therefore ,be introduced (ScHULZ, 1936). A method derived by Staudinger is based on the fact that the specific viscosity of a solution of chain molecules (i.e., the viscosity increase which is imposed upon the solvent by the solute), within a certain range of molecular weights, is approximately a linear function of the chain length. In addition to osmometry and viscometry we mention in particular Svedberg's ultracentrifuge for the determination of the degree of poly- merization of high polymer natural substances. X-ray analysis is not suitable for this purpose (see p. 99). 62 FUNDAMENTALS OF SUBMICROSCOPI C MORPHOLOGY I TiVBLE HOMOLOGOUS POLYMERIC SERIES OF CELLULOSE Degree of polymerization Chain length Mechanical properties Oligosaccharides y-cellulose HemicoUoid cellulose I-IO lO-IOO -50 A ^0-500 A Pulverizabl^ crystal powder Short-fibred /5-cellulose pulverizable powder Mesocolloid cellulose 100-500 500-2000 A Fibrous, strong a-cellulose (rayc^n) Native cellulose, 500-2000 0.25-1 /x Long-fibred, a-cellulose and more very strong (fibre cellulose) According to Staudinger, the experimental data available lead to the following conclusions regarding the molecule type of cellulose (Fig. 50). If some 10 glucose residues are linked together to form a chain, easily soluble cellulose products are obtained, which, owing to their particle length of 50 A, already exhibit slightly colloid properties. Compounds of this kind are known as degradation products of cellu- lose, termed cellodextrines or y-celluloses. If the number of chain links increases to 100, /9-celluloses are obtained which are soluble in 10% sodium hydroxide without swelling, to form viscous sols. Not before the degree of polymerization exceeds 100 and approaches 800 do we obtain the so-called a-celluloses, which are no longer attacked by 1% sodium hydroxide and which find application in the cellulose industry (rayon, cellophane). They slowly dissolve while swelling in 10% NaOH and yield viscous "gel solutions". Native cellulose has a still higher degree of polymerization; if dissolved in Schweizer's solution with complete exclusion of oxygen, a degree of polymeriza- tion of about 2000 for the fibre cellulose of linen, hemp, ramie and others can be calculated from the viscosity. The values determined from the viscosity can be checked osmometrically up to a degree of polymerization of about 1000 (Staudinger, 1936 a, b); beyond this limit extrapolation is carried out according to the linear viscosity STRUCTURE OF GELS 65 VIII (ACCORDING TO STAUDINGER, 1936b, 1937^) Capacity of film formation Solubility in 10% NaOH Viscosity in \ 1% SCHWEIZER solution Deviation from Hagen-Poiseuille law in I % solution 1 None Easily soluble Solution of low None Small without swelling Soluble without swelling viscosity Viscous solution None Large Slowly dissolved wuth swelling Viscous "gel solution" Small Very large Strong swelling almost insoluble Highly viscous "gel solution" Strong structural viscosity! rule. Whether this applies to the whole range from looo to 2000 chain links cannot be decided. Furthermore, it has been questioned whether chain molecules of such liighly polymeric substances can be completely dispersed in a micromolecular solvent at all (Lieser, 1940, 1 941). On the other hand, it is possible that native fibres contain still longer chains which may be degraded on dissolution in cuprammonium. The value of 2000 for the degree of polymerization of the fibre cellulose is, therefore, not reliable ; but it is the only value which can be determined at present experimentally and, for the time being, we must refer to it. Its magnitude is impressive enough, seeing that a degree of poly- merization of 2000 corresponds to a chain length of i /(, each glucose residue measuring 5 A. This means that the cellulose molecules have microscopic lengths. Nevertheless, they remain invisible because their thickness is amicroscopic. Chain molecules of a given structural type but different chain lengths are called a homologous polymeric series. The polyglucosans mentioned represent the polymeric homologues of the celluloses. In such a series the physical properties change with increasing molecular weight according to certain laws. Table VIII gives data for cellulose. Not only does the solubility decrease and the viscosity of the solutions increase, but the fibrous character and the capacity for film formation, which 64 FUNDAMENTALS OF S UBMIC RO SCOPIC MORPHOLOGY I are of particular importance in biology, become increasingly pro- nounced beyond a certain degree of polymerization. It is only from the low molecular weight members of a homologous polymeric series that uniform substances of definite molecular weight can be obtained by recrystallization, fractional precipitation, etc. In the members of higher molecular weight this is no longer possible. In the series of paraffins, in particular, it has been found that fraction- ation gives mixtures of substances of molecular weights which are only approximately equal. The determination of the degree of polymeri- zation therefore yields only an average value ; the actual chain lengths are spread more or less around this value according to the method of fractionation. Such mixtures are called polymer uniform substances ("'polymer einheitliche Stoffe") by Staudinger (1936b). Whether the high polymers occurring in nature are also polymer uniform, or whether life always builds chains of exactly the same length cannot be decided at present. Although the representatives of a homologous series behave quite differently from a physical point of view, they show the same, or at least a very similar chemical behaviour, in conformity with their uniform structure. For instance, the alcoholic OH-groups of all re- presentatives in Table VIII and, further, those of the polysaccharide molecules shown in Fig. 50 and even those of the polygalacturonic acid chains (Schneider and co-workers, 1936; Deuel, 1947b) can be etherified and esterified (methylated, acetylated, nitrated, etc.) with- out measurable change in the degree of polymerization. The polymer mixture formed in this way from the polymer uniform substance con- cerned has the same average chain length as the original material (it is an "analogous polymer", Staudinger, 1936b). On esterifica- tion, the cellulose chains lose their polar, hydrophilic properties, acquire a more homopolar lipophilic character and on account of their solubility in organic liquids are then more accessible to osmotic ex- periments. b. Structural Viscosity Anomalous flow. The four fractions of the series of homologous cellulose polymers yield colloid solutions of an entirely different nature. Staudinger divides them into "almost, meso-, hemi- and eu- -colloid" (Tabic VIII). In the two former cases the chain molecules in 2 STRUCTURE OF GELS 65 1% ScHWEiZER solution are completely solvated, i.e., completely surrounded by molecules of the solvent, and free to move as in real solutions. Their colloid character results merely from the fact that, the molecular length of solute molecules in one dimension being almost microscopic, they attract a large amount of solvent and thus increase the viscosity. Staudinger denotes this state as "sol solution". From a degree of polymerization of about loo onwards, however, a i% ScHWEiZER solution can no longer completely solvate all the chain molecules, and the solute molecules hamper each other's Brownian movement. They are not completely dissolved but are in a state intermediate between solid and liquid. At the highest degree of polymerization detectable, this interaction of the giant chains with 2000 links is so intensified, that the fibre cellulose dissolves very slowly. Solutions in which the chain molecules are hampered in their Brownian movement for want of solvent were called "gel solutions" by Staudinger (Staudinger and Sorkin, 1957b). There exists a reliable method, based on the phenomena of capillary flow, by which the concentration or particle size can be found at which the particles in a colloid solution begin to disturb each other, viz., Hagen- Poiseuille's law where q is the amount of Hquid flowing through a capillary of radius r in a time t under the influence of a pressure gradient p/1. In this for- mula the viscosity rj is independent of the pressure gradient p/1. This no longer applies when the colloid particles in the solution influence each other's motion. In this case the viscosity depends on the pressure gradient: r] = f(p/l), in the sense that the viscosity decreases with increasing pressure gradient. This can be explained by the fact that in these solutions the elastic properties of the solid substance are not completely eliminated, since the parcicles, instead of being fully dissolved, enter into some sort of relation with each other. With in- creasing pressure gradient in the capillary these elastic forces are progressively counteracted. For this reason, in colloid solutions with long chain molecules the chains which are originally present in a ran- dom and disorderly arrangement will be oriented parallel to the direc- tion of flow, and thus the forces resisting the flow which are respon- sible for the viscosity will be decreased. According to Table VIII, such 66 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY deviations from Hagen-Poiseuille's law are observed in the case of cellulose of polymerization degrees exceeding roo. Since the anomaly of flow is caused by the mutual positions of the colloid particles, it has been designated as structural viscosity (Ostwald, 1925; Philip- POFF, 1955). c. Gel Structure Gelj'rawe. If the coherence between the individual colloid particles becomes still more pronounced than in the "gel solution", gels are formed with a more or less fixed shape and distinctly elastic properties. Of course there exist all kinds of gradations from the gel solutions, in which the elastic coherence of the particles can only be proved by testing them for structural viscosity, and the real gels whose units are more or less fixed in their mutual positions. The gels that become liquid on shaking and soHdify again at rest, form a typical intermediate stage. This remarkable phenomenon is due to the same effect as the decreasing viscosity at increasing pressure; it is called thixotropy (cf. Freundlich, 1942). Colloid silicic acid and gelatin, for instance, can occur as thixotropic gels at suitable concentrations. If spherical colloid particles cluster together to a gel, the result is a rather compact gel and from Fig. 5 1 a it is obvious that such a structure can onlv be formed at a relatively high concent- ration of the solute. With sols containing long chain molecules, however, a fixed mutual position, i.e., a structure is attained much more easily. At concentrations as low as a few per cent, of a long chain high polymer, the chain molecules can combine into a loose meshwork, as represented by Fig. 51. Such a colloid already possesses a structure, although a very loose one, which may still easily undergo some plastic deformation. It also possesses a certain elasticity, because the places where the chains are interlaced can be regarded as fixed points. As will be shown, these can be due to various kinds of forces. Since in biolo- Fig. 51. Submicroscopic stnacture formation a) with spherical, ^) with filiform particles. 3 STRUCTURE OF GELS 67 gical colloids it is often difficult to decide upon their nature, I have suggested that they should be called by the neutral name oi points of attachment ot junctions ("Haftpunkte"), which is non-committal as to the kind of bonds involved (Frey-Wyssling, 1933 b; 1936 a). In Fig. 5 1 the junctions are marked by black dots. Obviously in a gel with chain molecules many fewer junctions are needed to build up a structure than in the case of spherical colloid particles. A gel built up by high polymer chains can therefore contain up to p/% of water CSeifriz, 1938) and yet possess a structure. This fact is very important to an understanding of protoplasm structure, since the water content of living matter is always surprisingly high. In Fig. 5 1 b further chains can be interwoven at will ; the number of junctions will then increase, and the result is a more solid gel structure. The plastic properties of the structure become less pronounced at the same rate, while the elastic properties increase. Thus the model of a gel structure projected here comprises all the states ranging from gels very rich in water to those very poor in water characteristic of active and dormant protoplasm. Limited sivelling. In the swelling process the absorbed medium pene- trates into the interstices available in the gel structure and widens the framework. It is clear that the permeating liquid should show chemical affinity for the chain molecules concerned. Thus the lipophilic mole- cules of rubber and polystyrene swell in benzene, which is lipophilic, while the hydrophilic cellulose swells in water. Whether in this process the framework of the gel goes to pieces, i.e., whether the gel is dissolved, depends on whether the junctions present can be disrupted. If the bonds are of the type of cohesive forces and the solvent present is capable of completely solvating the chain molecules, the gel structure may disintegrate and change into a gel solution in which the particles have greater mobility. This happens, for instance, in the swelling of fibre cellulose in cuprammonium. Limited swelling, therefore, always indicates that the chain molecules can only be solvated to a limited extent. Sometimes main valency bonds may be among the junctions. For example, as shown by Staudinger (1936a), polystyrene with a degree of polymerization of 1700 is soluble in benzene, but on the addition of traces of divinylbenzene (0.002%) it is converted into a product showing limited swelling in benzene (Fig. 5 2). In the same way chains 68 FUNDAMENTALS OF SUBMIC ROSCOPIC MORPHOLOGY CH=CH2 CH=CH! Divinyl- benzene -CH2—CH-CH2- CH - -CH2-CH -CH2—CH-CH2-CH — CH2-CH -CH2-CH-CH2- ■CH2-CH-CH2- ■CH2-CH-CH2- -CH2-CH-CH2— Fig. 52. Limited swelling of polystyrene (from Staudinger, 1936a). of methylcellulose can be interlinked by dicarbon acids (Tavel, 1939) or chains of polygalacturonic acid (pectic acid) by epioxides (Deuel, 1947a). When main valency bonds occur between the chain molecules, even the most suitable solvating medium is no longer capable of destroying the gel structure. Notwithstanding considerable swelling (e.g., a 30-fold increase in volume) the frame- work of the chains is 0^ ^^ preserved. It is possible [] that cellulose also con- V _ _ tains a few of such main valency bridges (see, for instance, Lieser, 1940; Meyer, 1940a; Pacsu, 1948). These would li- mit the swelling and would have to be de- graded chemically when cellulose is dissolved in cuprammonium. It can be said in general that limited swelling occurs when certain junctions of the gel frame (cohesive or main valency bonds) cannot be loosened. Concept of phase in gels. In the case of a sol one can (if necessary) speak of a "dispersed phase" distributed in a dispersing medium, although difficulties arise which have already been mentioned on p. 16. With sols containing chain molecules instead of colloid particles in the sense of the classical theory of dispersoids, to uphold the con- cept "phase" is decidedly wrong. For, according to the definitions in phase theory, separate molecules may not be characterized as phases. With gels the conditions are much the same. In a chain framework, it is incorrect to speak of a "dispersed phase", because regions with a thickness of molecular dimensions are not homogeneous phases, and the concept "dispersing medium" also becomes questionable. Consider a gel consisting of equal percentages of chains and water; a projection of the structure then gives the impression that the water is distributed as a "dispersoid" in closed compartments (Fig. 53a) whereas, conversely, in a cross-section of the gel the sections across the chain molecules appear as "dispersed" particles distributed in the liquid (Fig. 5 3b). In reality, however, neither of the two partners is STRUCTURE OF GELS 69 "dispersed" relatively to the other, for they both fill the available space continuously. A gel of chain molecules is therefore not a two-phase system but a single undivided phase. It is not only microscopically homogeneous and optically empty, but also homogeneous from a physico-chemical I - \ ^ / / \J b) Fig. 53. Gel framework, a) Projection, b) section across the frame- work. Areas to be compared encircled. point of view. As in the case of real solutions, if small volumes are considered, one always finds the same composition, with the sole difference that the volumes contain only parts of chains instead of whole molecules (Fig. 53b). Thus gels with a framework consisting of individual chain molecules are one-phase systems. As in the case of mesophases, this state deserves a nomenclature of its own. It will be designated as pseudophase, especially in view of the fact that often not all junction bonds are identical, so that the condition of homogeneity is not strictly satisfied. In concentrated gels the chain molecules show a strong tendency to orientate in parallel and to cluster in strands or rods. In such cases the parallel arrangement may become so pronounced that here and there the chain molecules combine to form a chain lattice. The length of the crystal lattice in the direction of the chain axis need not be the same as the length of these molecules ; the chains may protrude from the end planes of the crystalline rods (Gerngross, Herrmann and co- workers, 1930, 1932), continue further and eventually enter again into another region of lattice order, as has been indicated schematically in Fig. 54a. The more complete the average parallel arrangement of the chains, the greater the probability of the occurrence of crystals 70 FUNDAMENTALS OF SUBMIC ROSCOPIC MORPHOLOGY I (Fig. 54b). In this case the gel is no longer a one-phase system: the reo-ions of lattice order form a homogeneous phase in contrast to the pseudophase formed by the mixture of the unordered chains and the surrounding liquid. Hence, from a structural point of view there are two kinds of gels. iJ L_. a) b) Fig. 54. Ordered regions in a gel framework, a) Locally parallelized chain molecules, b) local formation of a crystal lattice. viz., I. one-phase gels whose framework consists of very long chain molecules interlinked at the junctions (pseudophases) and 2. two- phase gels with a crystalline and a non-crystalline (amorphous) phase. Instead of the fine chain framework of the one-phase gels, we then have a much coarser rod framework. Dispersion series. Having derived the structure of gels from the special form of the high molecular weight chain molecules — thus starting from below, that is from the amicroscopic domain — we shall now try to advance into the submicroscopic domain of gels from the macroscopic and microscopic regions. In colloid chemistry the concept of colloid particles is usually derived from macroscopic particles with the aid of a dispersion series. The particle size in this series decreases steadily to microscopic dimensions, ultimately declining to invisible submicroscopic dimensions. The final step in the direction of progressive dispersion leads from the colloid range to the ami- croscopic dispersions of true solutions (Table IX). When making a similar dispersion series for gels one must start from re ficu/ar instead of corpuscular systems. The frequency of such systems in biology is surprising; one comes to the conclusion that network systems of all dimensions are typical of living matter! STRUCTURE OF GELS 71 TABLE IX DISPERSION SERIES Order of magnitude of Corpuscular disperse systems Reticular disperse systems structural unit .•• • • • •"3-r^— K^ll ^^S Macroscopic Microscopic Submicroscopic Amicroscopic ^ Gravel i Sand Dust Clay Salt solution Liana undergrowth; veil of aerial roots Wad of threadlike algae VA gel, cell wall Methyl cellulose; cytoplasm(?) The entanglement of lianas in a virgin forest is a macroscopic net- work system (Fig. 5 5). A good example of fibre network is the veil of aerial roots of Cissus lianas in a tropical forest : thin filiform roots with a length of several metres hang slackly from the branches. They form, as it were, a fabric in the air, although none of these aerial roots have grown together. In moving air this entanglement of roots be- haves like a coherent mass because neighbouring filaments impede free movement. There are many other, still finer macroscopic net- work systems viz., skeletons of vascular bundles of leaves, succulent sprouts and fruits (Fig. 56), skeletons of sponges (especially not- iceable in silica sponges), spongy parts of bones, etc. An excellent example of a meshwork with microscopic elements is macerated skin (Fig. 57); also latex tube systems of the latex plants. When algae threads are fished out of a pool, we are astonished to find how they cling together in a tangled skein, although every thread is an individual in itself. Here the junction bonds, which are hypothetic in the case of gels, can actually be observed under the microscope, for at all points where two threads cross, they stick together (cf. Fig. 51b, p. 66). The number of these junctions is so great that a wad of algae like this is even slightly elastic when compressed. We penetrate into the submicroscopic domain by making the threads so thin as to become invisible under the ordinary microscope, thus obtaining gels. Until recently their structural principles had to be found out by indirect means (Frey-Wyssling, 1938). Nowadays, 72 FUNDAMENTALS OF SUBMICROSCOPI C MORPHOLOGY Fig. 5 5 500 fJ- Fig. 56 i^ -I Fig- 57 Fig. 58 Reticular structures of different scales Fig. 5 5 . Coarse macroscopic reticular structure : liana brush in a virgin forest photographed skyward. Fig. 56. Macroscopic reticular structure: fascicular skeleton of a Luffa fruit (vegetable sponge). Fig. 57. Microscopic reticular structure: network of collagen fibres in cow's skin (Kuntzel, 1941). Fig. 58. Submicroscopic reticular structure: ultra-structure of coagulated blood fibrin (from WoLPERS and Ruska, 1939). 3 STRUCTURE OF GELS 73 however, the reticular structure of gels can be photographed directly in the electron microscope (Fig. 58). As will be explained in the next paragraph, the submicroscopic strands or strings which form the gel frame will be designated as micellar ttrands. Thus the submicroscopic gel structure is a micellar framework. The transition into the amicroscopic domain is of particular im- portance. Whereas corpuscular disperse systems in this case become real solutions and are no longer accessible to colloid chemical- methods of research, reticular systems remain colloids even if the thickness of the strands of their framework is reduced to amicroscopic dimensions, i.e., to the cross-section of a single molecule. Thus in net- work systems there is no lower Umit to the colloid domain ; they re- main gels irrespective of whether their network is submicroscopic or amicroscopic. Examples are the polystyrene gels mentioned on p. 67, (Fig. 52, p. 68), or the methyl cellulose gels prepared by Tavel (1939) with the aid of oxalyl chloride, or pectin gels prepared by Deuel (1947a) with ethylene oxide. In these cases the strings of the network are chain molecules and the gel structure is a fine molecular frameivork. Comparison of corpuscular and reticular systems. The properties of net- work gels differ in principle from those of sols with their corpuscular dispersed particles. This is clearly demonstrated by Table X. Whereas a liquid capable of solvating a substance will disperse corpuscular colloids, reticular colloids remain a coherent mass into which the solvating medium can penetrate to a certain extent only (limited swelling). In this case the dispersing medium would be better characterized as an imbibition medium (see p. 81 and 84), since the colloid substance is not dispersed into separate particles. In the coacervation of sols an equilibrium liquid poor in colloid and the coacervate layer rich in colloid are formed (Fig. 21, p. 22). In reticular coacervates, however, the equilibrium liquid contains no colloid, be- cause the latter is insoluble in the reticular state. For example, after gelation of a gelatin solution, no gelatin is found in the supernatant liquid (cf. p. 21). In reticular colloids the mutual position of their submicroscopic elements is fixed, so that a structure results. It follows that gels possess a certain elasticity, although often only slight, indicating that the forces acting in the junction bonds are weak. Typically intermediate between gels and sols are gel solutions, whose particles impede each 74 FUNDAMENTALS OF SU BM I C RO SC OPI C MORPHOLOGY I TABLE X COMPARISON OF CORPUSCULAR AND RETICULAR COLLOIDS IN THE SOLVATED STATE Corpuscular colloids Intermediate state Reticular colloids Colloid state Sols Gel solution Gels Colloid portion Individual particles Particles are impeding Coherent structure (micelles or macro- each other's motion (micellar or macro- U3 molecules) molecular frame- _u work) Solvating liquid Dispersing medium Imbibition medium Equilibrium liquid Dispersing medium Imbibition medium 1— « incoacervation -f colloid portion fjee from colloid Elasticity Inelastic Structural viscosity Elastic Structure Structureless Short-range order Structured Ultramicroscope Demonstration of Gel frame is optically j: particles empty Ultracentrifuge Sedimentation Syneresis CO Ultrafiltration Particle size Pore size U4 Dialysis i With the aid of > Without membrane O W3 Donnan equilibrium ^ membranes -X3 C Osmotic laws Hold good Disturbed Do not apply, because -C gel is insoluble Kinetic migration Mutual diffusion Permeation Electric migration Electrophoresis Electrosmosis ■|i Dilution, swelling Unlimited dilution Unlimited swelling Limited swelling Disturbance of Precipitation "Hardening" irbanc ilibri. stability (flocculation. (tanning, fixation) coagulation) ■ 22 g" Separation into Two coexisting Usually vacuolization C '^ different phases layers Other's free movements. These gel solutions, therefore, show structural viscosity, demonstrating the existence of some structure when de- formed. The difference between corpuscular and reticular systems is particu- larly apparent when testing the apphcability of the methods of research developed by colloid chemistry. As a result of the close packing of the micellar strands, all gels are optically empty in the ultramicroscope. In the centrifuge no definite sedimentation equilibrium is established; 5 STRUCTURE OF GELS 75 some of the imbibing medium is simply pressed out of the gel (syn- eresis). As the gels do not contain individual particles, ultrafiltration cannot be used as a method to discover whether they contain sub- microscopic or amicroscopic structural elements. It gives some in- formation, however, about the approximate pore size of the network structure, since on account of its structure each reticular colloid re- presents an ultrafilter, provided it possesses the firmness needed to resist the fihration pressures applied. In all other methods of research mentioned in Table X the contrast between the movable particles of sols and the immovable frame of the gels finds expression. In dialysis and in the study of Donnan equilibria, amicroscopic particles are removed by diffusion through a membrane which is impermeable to colloid particles. In the case of insoluble gels no membrane is needed, because the colloid portion is itself immovable (see p. 202). For the same reason, the osmotic laws are not applicable to gels, whereas in true sols, where the individual particles are completely in- dependent, they allow of a determination of the number (and therefore the weight) of the particles. Finally, when concentration gradients or potential gradients are applied to gels, the amicroscopic particles diffuse through the gel frame (permeation), or the imbibition liquid migrates through the electrically charged network (electrosmosis). Similarly when the equilibrium in a colloid system is disturbed, the behaviour of gels and sols is fundamentally different. Sols can be diluted by the solvating liquid, whereas in true gels only limited swell- ing occurs. In sols the disturbance of stability factors (hydration and charge) may lead to flocculation or coagulation. In contrast to what is commonly asserted, gels do not coagulate, they are "hardened". In technology this is denoted by tanning and in cytology by fixation. Separation of sols results in two microscopically uniform "phases" (Fig. 15, 21, p. .17, 22), whereas in gels the separated drops usually cannot unite and give rise to vacuolization in the originally micro- scopically uniform system (Fig. 23, p. 23). The concepts of limited swelling, fixation and vacuolization, which are mentioned at the bot- tom of the last column in Table X, are familiar to all cytologists and we need waste no time on the question as to which colloids are of the first importance in microscopic and submicroscopic morphology. Indeed, the number of colloid systems in biology, whose nature has been ascertained successfully by means of the methods of research 76 FUNDAMENTALS OF SUBMI C RO S COP I C MORPHOLOGY I developed for corpuscular dispersoids, is very small (blood, mil k, serum, suspensions of micro-organisms and viruses). No conclusive information could be derived by these methods on the fine structur e of the protoplasm. The very terminology of the theory of dispersoid s, which assumes dispersed particles in a dispersing medium, is unsuitable. True, the introduction of dijJorM, i.e., strongly anisodiametric par- ticles, accounts to a certain extent for the properties known to modern macromolecular chemistry (Manegold, 1941). The older technical terms of Nageli (Nageli and Schwendener, 1877) are much better adapted to the needs of biologists working with gels. Nageli's ideas can be applied to our present concept of gel structure. To that end let us first give a precise definition of the micellar concept, to which unfortunately various meanings have been attached in colloid science. d. Micellar Theory The concept of the micelle. C. Nageli was the first to develop a well-founded theory on the structure of hydrogels, which he designated as organi-:^ed substances. Starting from double refraction, anisotropy of swelling and layer structure of grains of starch (18 5 8, new edition 1928) and of cell membranes, he made the assumption that these substances consist of long, submicro- scopic particles, supermolecular in character and of crystalline structure. Such a particle was called a micelle (diminutive of the latin mica = a crumb or little bit). Later, Nageli extended his theory to solutions. He stated that, when a gel is dissolved, the micelles are maintained as units and give a micellar solution. As a result of this transference of the micellar concept from solid gels to solutions, this concept is used in the literature in various meanings, as has been pointed out by several authors: Zsigmondy (1921), Ambronn and Frey (1926, p. 152). Whereas the biologists, in particular Ambronn's school (Frey, 1926a, 1928a, b) and also Schmidt (1934), adhere to the original definition which indicates the: form and crystallinity of the particles, the meaning attached to micelles by colloid chemists is as a rule simply that of dispersed particles in a colloid solution, stressing in particular their electrical charge, without heeding their form and structure. In the latter case, therefore, it represents an overall concept which may embrace all possibilities such as primary particles (monones), secondary particles (polyones), associate^ (^-g-. i^^ soaps), etc., including their charges and solvation layers. As a result of this situation, the origin of this term is scarcely known in colloid chemistry. This led to what Nageli objected to in a discussion of Pfeffer's terminology in the famous "Osmotische Unter- suchungen" (1877). Nageli says (new edit. 1928, p. 70/71): "Pfeffer uses the general expression 'tagma' for molecular compound, observing that in chemistry one would hesitate to introduce the term micelle, which is re- 3 STRUCTURE OF GELS 77 miniscent of cell. It seems, therefore, that the etymological error is made : to believe that we are deaUng with a barbaric composition of "cellula" and an unknown word beginning with "mi", in much the same way as the word aldehyde is formed." By optical means, Ambronn has definetely established the existence of o) Fig. 59. Former conception of the micellar structure: a) from Nageli and Schwendener (1877), b) from Seifriz (1929) and K. H. Meyer (1930). long, submicroscopic particles in gels such as celloidin, denitrated cellulose, celluloid, gelatin, aluminium oxide fibres (see p. 82). These particles often showed an intrinsic double refraction which could only be explained by assuming crystaUine particles (Ambronn, 191 6/1 7). The existence of crystaUine micelles in chitin (Mohring, 1922), in muscle fibres (Stubel, 1923) and in vegetable cell walls (Frey, 1926b) was demonstrated by means of the same methods. At about the same time the crystalline nature of many colloid particles, for example gold sols, cellulose and many other colloids, was estabhshed by the X-ray method (Scherrer, 1920). Nageli's micellar theory was taken up by Meyer and Mark (1930) and propagated by them among chemists in an almost unaltered form, after having been nursed for a long time in its original form by a few biologists. This is obvious from a comparison of Nageli and Schwendener's scheme (1877) and the model of fibre struc- ture given by Seifriz (1929) and K. H. Meyer (1930): Fig. 59b. In -}% FUNDAMENTALS OF S U BM I C RO S COPI C MORPHOLOGY I Nageli's scheme (Fig. 59a), two intermicellar substances are drawn between the micelles; one of these substances may be eliminated. What is new in Fio-. 59b is the determination of the inner structure of the micelles; for the rest, however, there is complete agreement with Fig. 59a. The micelles were considered as disperse phase, surrounded by intermicellar spaces which are accessible to the dispersing medium. To account for the coherence of the crystalline micelles in a solid framework, special micellar forces had to be assumed. Meyer and Mark considered these to be large cohesive forces which, in cellulose for instance, are additively composed of the molar cohesions of the numerous OH-groups. However, since these same forces act intramicellarly as lattice forces, it was difficult to see what the difference might be between the forces responsible for the />//ramicellar coherence of the chain molecules in a crystal lattice and the /«/^rmicellar "micellar forces". According to Nageli, when a gel is dissolved, the micelles are dispersed, and the sol contains independent crystals. This point of view has often been adopted by others, in particular for cellulose sols, although such solutions do not give X-ray diagrams (e.g.,HERZOG, 1927). According to Staudinger (1932), the high polymer natural substances are dissolved as separate chain molecules instead of crystalline particles. At present, therefore, only crystal- Une suspensoid colloids such as gold-, vanadium pentoxide-, ferric oxide sols, etc. can be claimed to be micellar solutions in Nageli's sense; they show mostly a strong birefringence of flow and partly also X-ray inter- ferences. In the case of gels, our recent knowledge of the structure of high polymers raises further objections to Nageli's concept of micelle, for it is found that the chain molecules are much longer than the crystalline regions (Fig, 54, p. 70). It follows that the micelles, instead of possessing individual character as assumed by Nageli, have grown together and are to a certain extent absorbed in the gel structure. Nowadays they can no longer be considered to be substantial (not even conditionally substantial) particles (Frey-Wyssling, 1936a, c; Kratky and Mark, 1937). They consist of well- ordered chain molecules, which protrude from the crystalline into the amor- phous regions and perhaps take part again in other ordered lattice regions. We conclude that i . there are sols containing chain molecules which are more or less independent, rather than micelles in Nageli's sense, and 2. that the micelles in gels do not represent independent crystallites but at best can be described as lattice regions. Taking into account the constant danger of confusion with the colloid chemical concept, which by micelle means an electrically charged instead of a crystalline particle, it would perhaps be better in our considerations to give up the concept micelle. If one wishes to use it nevertheless, one should not assign any special significance to this concept, but simply use it in the sense of supermolecular colloid particle. This would exclude all possibility of confusion. One would then have to distinguish between two different kinds of colloid particles: i. super- 5 STRUCTURE OF GELS 79 molecular micelles consisting of many molecules, and 2. macromolecular molecules of submicroscopic dimensions. However, since a well-founded terminology for sols does already exist, the micellar theory will be confined to gels, as originally intended by Nageli. Nomenclature. Although the assumption of independent micelles in gels has proved to be erroneous, Nageli's work contains a great many other ideas on the structure of gels which have been shown to be quite correct. I quote the following paragraph, for instance, (new edition 1928, p. 76/77): "Die Micelle vereinigen sich . . . zu Verbanden . . ., indem sie sich beliebig, bald baumartig, bald mehr netzartig anein- ander hangen. Diese unregelmiissigen Verbande . . . bilden eine ste- hende Gallerte". Elsewhere he speaks of "Micellar-Reihen, in denen die Micelle miteinander verwachsen sind". Although at the time the existence of chain molecules was not even suspected, he has given a description of gel structure which is essentially correct. To current biology the main concern is, not whether living matter and its derivatives contain or do not contain crystalline regions, but rather whether the particles are independent of each other, as presumed in classical colloid chemistry or in the theory of dispersions, or whether they are united in a framework (however weak), and thus provide the colloid with a structure. Consequently, contrary to structureless dispersoids, gels are in need of an appropriate terminology. It is tempting to make up for this deficiency by creating new names^ However, one does not always render science a service by doing so, and it is perhaps preferable in this case to use old well-tried expressions adapted to modern experimental results by new definitions. Following Nageli, the frame substance will be designated as micellar portion and the interstitial substance as intermicellar portion of a gel. In those cases where the micellar structure consists of coarse beams or joists, which are partly crystalline and therefore homogeneous, one can also speak of micellar phase and intermicellar phase. There is no danger that this new definition will again give rise to confusion, for the concept intermicellar is used in exactly the same sense as hitherto in the literature of the subject, and the concept micellar is only changed so as to apply not exclusively to the crystalline regions of a framework, but to the framework as a whole. This solves the ^ Pfeiffer (1941b, i942a) designates the theory of fine-structure as leptonics and the invisible structural units as leptones (from Ac-tto'c = fine, small). 8o rUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I difficulty that gels whose framework units consist of only a few parallel chain molecules do not answer to Nageli's original definition, because a small number of chain molecules are not capable of forming a crystal lattice. With still finer strands of the gel structure, it is true, it ultimately consists only of chain molecules and the micellar frame- work has changed into a molecular frameivork^ as has been pointed out on page 73. Just as in the transition from colloid to molecular disper- sions, there also exist transitional forms between (a) gels with micellar strands and (b) gels with chain molecules as structural units. In gels with a molecular framework the particle size of the two components of the system are not similar as in the case of solvent and solute molecules in a true solution. In principle they remain different in a morphological sense as framework and interstitial substance. In the case of micellar systems possessing strands with a thickness of several molecules, a distinction should be made between processes which occur in the meshes of the network (/>^/^r/;//V«'//<«r) and those occur- ring inside the beams of the frame, i.e., in the crystal lattice {intra- micellar). In the same sense the expression "intramicellar" is used for cation exchange inside layer lattices (Wiegner, 1935 ; Bottini, 1937). With the aid of the concepts "micellar", "intermicellar" and "intra- micellar", all processes occurring in gel structures can be described unambiguously. By a relatively sUght change in concepts we thus preserve a nomenclature which has done good service for 90 years, and renders honour toNAGELi,who laid the foundations of the research on biological gels. In Table X we have recapitulated the most important points which, according to our definition, distinguish the reticular gel from its counterpart, the corpuscular sol. As in the case of dispersoids (Table II, p. 16), the components of a gel can occur as gases, hquids or solids, with the restriction, however, that the micellar component must always be solid (Table XI). If the intermicellar substance is a gas or a liquid, we have to deal with network structures or capillary structures. If it is a solid, however, solidified gels result, showing clearly in contrast to dispersoids that the two components are completely equivalent as regards the arrangement in space. The micellar structure is determined by the micellar strands, by the type of bonds between them and by the intermicellar substance. For a given type of micellar units, however, the gels can be built up with STRUCTURE OF GELS 8i TABLE XI (Compare Table II) RETICULAR SYSTEMS (ACCORDING TO FREY-WYSSLING, 1 93 yd) Imbibition medium Micellar frame Structures Solid Liquid Gas Solid Solid Solid i Composite solid } Gel structure ^ Capillary structure various different possibilities of orientation. This determines the micellar texture, which gives information about the arrangement of the structural elements in the gel, in contrast to the micellar structure, which characterizes the fine-structure in general. Definitions. To sum up, we give the following survey: By structure^ we mean the fixed mutual positions of the submicro- scopic or amicroscopic morphological units ; by texture, the special arrangement and distribution of such structural units. = colloids with freely moving particles = colloids with a gel frame = molecular colloid particles = supermolecular colloid particles, most often packets of chain molecules in parallel arrangement == amicroscopic structure of intertwined cliain molecules = submicroscopic structure of coherent micellar strands = substances in the interstices of a molecular framework = substances in the interstices of a micellar frame- work = processes occurring between the strands of a gel frame = processes occurring inside the strands of a gel frame = fine-structure of gels in general = arrangement of the structural units in particular ^ Not only crystalline but also amorphous solid phases possess a structure. For, in amorphous glasses (Bussem and Weyl, 1936) and also in isotropic gels the structural elements are bound together elastically in fixed mutual positions, notwithstanding the lack of order. We must therefore in principle attribute a structure to ?11 solid states of matter. corpuscular colloids reticular colloids macromolecules micelles molecular framework micellar framework interstitial substances intermicellar substances intermicellar processes intramicellar processes micellar structure micellar texture 8i FUNDAMENTALS OF SUBMICROSCOPI C MORPHOLOGY E § 4. Studies in Gels The colloid chemical methods of investigation which have proved so- successful in the elucidation of the nature of sols have only a limited applicability to gels (compare the discussion of Table X, p. 75). Gels must therefore be investigated by different means. Of these we shall only discuss those which are of special importance to the investigation of cytological objects, leaving others, such as are of interest, e.g., in the technical testing of gels, out of account. For lack of space the methods of investigation will not be treated in great detail ; we shall only deal with the principles of these methods and the problems which they can solve. a. Polarisation Microscopy Theory of composite bodies. The texture of gels can be explored by optical means if two conditions are fulfilled. In the first place the "//="/ i"//="< ni-n,f_ nj, = n^ Fig. 60. a) Rodlet composite body, ny big, Ha small refractive index, b) Layer or platelet composite body, Oy big, Hq small refractive index. strands of the framework must be separated from the intermicellar space by definite phase boundaries, and secondly they may not be oriented at random but must show a certain tendency to orientation in a given direction in space, Wiener (191 2) has calculated theoretic- 4 STUDIES IN GELS 85 ally the optical effects occurring in these systems. In this calculation one must assume greatly idealized textures with, for instance, parallel circular cjdinders or parallel planes (Fig. 6oa and b). Such aggregates, meeting the mathematical requirements, are designated as "composite bodies" (German: Mischkorper). As the structural units (cylinders or planes) are not bound together, they do not possess a micellar structure in our sense. One can imagine, however, that a gel is formed out of such an idealized composite body if the structtiral units are somehow anastomosed with each other. This does not affect the general character of the optical effects, but it is obvious that quantitative calculations according to Wiener's formulae cannot give very accurate results for gels with a micellar structure, since the geometrical conditions for an accurate mathematical treatment of the problem are not satisfied. The rods or layers of the composite body must be supposed to be optically isotropic. It then follows from theory that the behaviour of the composite body with respect to polarized light depends on the direction of its vibration, i.e., such a body is anisotropic, provided the diameter of the cyHnders and the distances between the cylinders or layers are small compared with the wavelength of the light. It should be borne in mind that by "small" we do not mean arbitrarily small, as the structural units should possess true phase boundaries. Single chain molecules, for instance, cannot act as structural units in a composite body. Optical anisotropy can manifest itself in three different ways : 1 . Birefringence. The refractive power (n^y) for directions parallel to the axis of the composite body is different from that perpendicular to it (nj^) (Fig. 60a, b) so that, in polarized light, interference colours occur as in doubly refracting crystals. 2. Anisotropic absorption (dichroism). In coloured composite bodies, absorption is different parallel (k^^) and perpendicular (kj^) to the axis ; they therefore show different colours depending on their position with respect to the plane of oscillation of linearly polarized light (Fig. 6i). 3. Anisotropic diffraction. Transmitted light is differently diffracted in different directions; the typical gloss of silk, for instance, must be attributed to this effect. The composite bodies possess a very typical characteristic: their anisotropy is not constant but is a function of the properties of the sub- stance enclosed between the particles, which in microscopy we de- 84 FUNDAMENTALS OF S U BM I C ROSC OP I C MORPHOLOGY I sign2itc2.smoufitmg\iqmd, or better as imbibition liquid. Hence the double refraction changes with varying refractive index n of the mounting liquid. For this reason the double refraction of such composite bodies s Fig. 6i. Dichroism of bast fibres (SE vibration plane of polarizer) from Frey (1927b). a) Stained with chlorozinc-iodine : black/colourless; b) stained with gold: green (marked with little lines)/claret (dotted). differs essentially from the double refraction of crystals, which re- presents a constant characteristic of the crystal. Fig. 62 shows the changes in double refraction observed when epidermal hairs of incinerated barley awns are mounted successively in air (n = i.oo), water (n = 1.33), alcohol (n = 1.36), xylene (n = 1.49), benzene (n = 1.50), Canada balsam (n = 1.54), mono bromo naph- thalene (n = 1.66), potassium mercury iodide (n = 1.72). The variation of the birefringence with the refractive index no of the imbibition liquid obeys a hyper- bolic law. The double refraction is zero when n^ = n^ (n^ = refractive index of micellar component). Composite bodies with rodlet texture are optically positive, those with layer texture are negative. Wiener's formula for the rodlet bire- fringence runs : in A 1200 900 600 300 \ \ / \ / ^ \ / \ ^^.^ 10 1.2 14 1.6 1.B 02 Fig. 62. Curve of rodlet birefrin- gence of the epidermal hairs of in- cinerated barley awns (from Frey, 1926b). Abscissa: refractive index ng of the imbibition liquid. Or- dinate: retardation yk in A units. n n ■1 (^ i)n| + ^2^1^ Here n^^ represents the extraordinary refractive index (parallel to the axis of the composite body) and n_^ the ordinary index (perpendicular to the axis), n^ the refractive index 4 STUDIES IN GELS 85 of the isotropic rods and n^ that of the imbibition hquid; d^ and 62 are the volume fractions of the two components (^1 + ^2 = i)- Clearly, n?, — nl is a measure of the double refraction n^^ — n^. The formula shows how this double refraction depends on the refractive index n.^ of the imbibition medium. It is zero when n^ = n^, and positive for all other values of n.,, because the numerator contains the square of nf — n|. In other words, the rodlet birefringence is always positive : n,, > nl- Since in birefringent objects the larger index is denoted by ny and the smaller one by na, it follows that n^^ = ny and n^ = Ua. Conversely, in composite bodies with layer texture and negative bi- refringence we have nj_ = riy and n^^ = na. It is significant that besides the volume fractions 6^ and 62 no quantities depending on the dimensions of the rods occur in the equa- tion. The double refraction is independent of the thickness of the rods. This is of particular importance to the study of submicroscopic textures, as long as the size of the structural units is not known. The double refraction of the composite bodies has been termed form birefringence (Frey, 1924), because its nature depends on the form of the textural elements of the solid phase. The curves of form birefringence are therefore used to examine whether intermicellar spaces occur in a material and to decide whether the micellar phase has the form of rods or platelets. Usually one does not measure the birefringence n^^ — n^ itself, since this depends on the variable thickness d of the swollen gel according to the formula njj — nj_ = yA/d, but simply the retardation yX, where y is the so-called phase difference and A the wavelength of the light. The introduction of this method of research into colloid optics is due to Ambronn. Measurement of tjje birefringence. The basic formula for birefringence can be simplified by introducing the notations Zln for n^^ — nj^ and r for the retardation or path difference yX. This gives /In = r/d, which shows clearly the linear dependence of the retardation on the thickness d of the object, because zln for a given object in a given medium is constant. The retardation F is measured by a compensator. This is a crystalline lamella (quartz, gypsum, calcite) with known double refraction An which 86 FUNDAMENTALS OF SUBMI C ROSCOPI C MORPHOLOGY I is inserted into the polarizing microscope. It is in the form of a sliding wedo-e, or a flat plate which can be tilted so that its thickness d is variable. Since the light oscillating parallel to the direction of the minor refractive index of a double refracting specimen passes faster through the object than the beam oscillating in the perpendicular direction parallel to the major refractive index, a path difference of these two beams results, which causes the interference colours observed in the polarizing microscope. This retardation can be diminished if the direction of the major refractive index of the specimen is oriented parallel to the minor index of the compensator. By varying the thickness of the compensator, the retardation of the specimen can be counterbalanced, until the colours disappear completely. Then the double refraction is compensated and the value of F can be read from the compensator. For delicate measurements there are compensators which permit determination of the phase difference y of the two beams. Then the readings must be multiplied by the wavelength A of the monochromatic light used, or by A = 550 m/x for white light. The formula mentioned above applies to objects bounded by two parallel planes as, e.g., in microtome sections, where d corresponds to the thickness of the section. Many biological objects, however, (myelin tubes, myelin sheath of the nerves, fibres with narrow lumen, etc.) occur in the form of hollow cylinders. In this case the thickness becomes greater with increasing distance from the edge, and according- ly the path difference increases. The phenomena are particularly com- plicated when the optical axis is not parallel to the axis of the cylinders as in fibres, but perpendicular to the cylinder axis, as is the case of myelin objects. The birefringence An. may then be calculated from a formula of Bear and Schmitt (1936) if the largest possible path difference /^(max) is measured. This formula runs : (d, + zda) arc cos [(d^ + 2d2)/3di] where d^ represents the diameter and d., the inner diameter of the hollow cylinder. A similar problem occurs in the determination of the double re- fraction of objects with spherite texture and radially oriented optical axis Ce.g., grains of starch). In this case the double refraction is 1 (max) An 1. 122 r where r is the radius of the spherites (Frey-Wyssling, 1940b). Bear and Schmitt's formula should yield this value for a solid cvHnder, 4 STUDIES IN GELS 87 where d.^ = o. This is not so, however, because empirical data con- cerning the position of the maximum retardation /"(max) in a myelin tube have been mixed up with the optical theory (Schmitt and Bear, ^937)- S2gn of the double refraction. The sign or character of the double re- fraction is a very important datum for the textural analysis of gels. A micellar texture is called optically positive if, as mentioned before, n,, — nj_ has a positive value. If, on the other hand, n„ — nj^ is smaller than zero, the double refraction is negative. The refractive index n,, always refers to a direction which in some way or other is of a special character : direction of the orientation in the composite l)odies mentioned, direction of growth, direction of pressure or ten- sion, direction of flow, special morphological direction, and so on. In fibres and threads, for example, the fibre axis is the reference axis, in cross-sections of parenchyma cells the tangential direction, in spherites the radial direction. The character of the birefringence of gels is indicated by the so- ■caUed index ellipsoid, the long axis of which corresponds to the larger index ny, while the short axis corresponds to the smaller index na. The direction of ny is determined by comparison with a selenite 1 [ I J - 1 r-" - . L J J-L o) + b) \ I ii I -i- CI^ d) + + f) T-ig. 63. Optical character of gels. Reference axis marked by a dotted line, a) Rodlet ■composite body, b) layer composite body, c) thread of gum arable, d) thread of cherry gum, ;) section across a vegetable parenchyma cell (reference axis = tangential direction), /) starch grain (reference axis = radial direction). plate (see Ambronn and Frey, 1926). The orientation of the index ■ellipsoid and the direction to which the double refraction refers have heen drawn in Fig. 63, Many gels are isotropic when observed in the direction of the reference axis; they are uniaxial in the crystallographic 88 FUNDAMENTALS OF SUBMIC RO SCOPIC MORPHOLOGY I sense, and the definition of optically positive and negative is in com- plete conformity with the terminology customary in mineralogy. In those cases, however, where the object shows anisotropic behaviour towards Hght incident along the reference axis, crystal optics use other definitions to describe the optical character, and the customary terminology in gels is no longer identical with that in crystal optics. Whenever there exists a direction of isotropy, this should be chosen as reference axis. Systematics of double refraction. In most cases the micellar texture itself is birefringent, because the chain molecules constituting the strands of the structure are themselves anisotropic. This kind of optical aniso- tropy is called intrinsic double refraction. In this case the double refrac- tion of the gel cannot be reduced to zero by changing the refractive index n^ of the imbibition Hquid; there is a residual double refraction in the minimum of the curve for form birefringence : the intrinsic double refraction of the substance. In all cases examined so far, the micellar strands behave like optically uniaxial systems, or at any rate at a first approximation. They possess, therefore, two principal refractive indices, designated by n^ (extraordinary index along the fibre axis) and no (ordinary index perpendicular to the fibre axis). The intrinsic double refraction is accordingly ng — no. As a rule it is positive, but sometimes turns out to be negative. In those cases where the intrinsic double refraction is different from zero, the refractive index n^ in Wiener's formula is to be replaced by the average value \ (ng + no) or, better still, by \ (ng + 2no). Both types of form birefringence (positive composite bodies with rodlet texture and negative composite bodies with layer texture) may be combined with the three possibilities, positive, negative and zero intrinsic double refraction. On the whole one can, therefore, distinguish between six types of double refraction (Frey, 1924). Both the form and the intrinsic birefringence can be attributed to the structure of the object, but the intrinsic double refraction is caused by the much finer structure of the crystal lattice, whereas the form birefringence results from the coarser colloid structure. Hence the latter is as a rule smaller than the former. The intrinsic and the form double refraction are both due to morphological properties, in contrast to the phenomenon of incidental double refraction^ which becomes apparent when solid objects are sub- 4 STUDIES IN GELS 89 ject to tensions or pressures; the designation is, therefore, double refraction due to tension or tension double refraction. This phenomenon accompanies elastic deformation (photo-elastic effect), and elastic de- formability is a condition for its occurrence. Since, according to definition, gels actually do possess this property (Table X, p. 74), effects of this kind are to be expected in gels exposed to stress. The tension double refraction is usually positive with respect to the axis of deformation, while that due to pressure is usually negative. The effect is most pronounced in isotropic gels (e.g., strain-free gelatin), but is of course also observed in gels which are anisotropic by nature if these are exposed to tensions, in which case it is superposed on the pre-existing textural and intrinsic double refraction. On removal of the stress, the tension double refraction must disappear, as with every really elastic phenomenon. If it does not, the object has been plastically deformed. The photo-elastic effect is due to the deformation of elec- tron orbits in the material concerned ; the distances between the atoms in this material are slightly increased or decreased. In cubic crystal lattices insignificant changes in atomic distances cause considerable optical anisotropy (Wiener, 1926b). Orientation double refraction. The junction bonds in a gel being seldom very strong, they easily yield to the forces apphed. The elastic deforma- tion is then followed by a re-orientation of the micellar strands, thus intensifying the intrinsic and textural birefringence of the gel. For this reason the optical phenomena in gels exposed to stress are often very comphcated. The difference between the double refraction due to tension and that due to orientation is most obvious when these phenomena are different in sign, as for example in the basic experi- ments of Ambronn (1889) with cherry gum. For, when cb^rry gum is stretched, the transient, weakly positive double refraction resulting from the tension is followed by a negative double refraction due to the orientation of the micellar texture. D'stribution of orientations. In a stretched gel, the directions of the micellar units are spread about the reference axis according to a com- plicated distribution function (Kratky, 1933, 1938). The majority of micellar strands enclose small angles with the direction of the stretch, and only few of them enclose large angles with this direction. The distribution function depends on the degree of stretch. If this strain is unknown, however, an idealized scheme of the distribution 90 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I can be made by assuming that within a certain angle all possible orientations about the reference axis occur with equal frequency. The assembly of orientation then forms a sector (in a plane) or a cone (in space), whose vertical angle a can be computed from the double re- fraction of the gel when the intrinsic double refraction ng — n^, of the micellar strands is known, provided that by judicious choice of the imbibition liquid negligible form birefringence is assured. The angle of scattering a is then given by the following simple relation (Frey- Wyssling, 1945): r ■ ■ , A , . sin 2a for scattering in a plane Zln = (ng — nj for scattering in space An = (n^ — nj 2a COS a + cos'^a For example, the space angle in cellophane paper, referred to the preference direction, imposed by the manufacturing stress, was found to be 71°. 5. The anisotropy of cellophane is, therefore, rather strong, for the angle of scattering corresponding to the isotropic state, i.e., completely uniform distribution, would have been 90°. The micellar strands with their numerous orientations in space may be replaced by a gel in which only a single orientation occurs. This orientation angle is called the average orientation angle am- With the assumptions made by us am becomes ^a, as shown by Fig. 64a. The orientation of the strands in a micellar texture can be brought about by a variety of means other than tension or pressure, e.g., by drying or freezing a gel (Ambronn, 1891 ; Ullrich, 1941); the strings or strands of the frame are then shifted into more or less parallel positions. Birefringence of flow. The best-defined orientation, however, is that in a field of flow, if one succeeds in liquefying the gel to a sol by re- leasing the junctions. If such a solution is subject to flow, the colloid rodlets are turned parallel at all points where a velocity gradient exists. A well-defined velocity gradient can be obtained by introducing the sol into a narrow gap (width below 14 m^i) between a fixed hollow cylinder and a revolving inner cylinder (Signer, 1930, 1933; Boehm, 1939; Frey-Wyssling and Weber, 1941). When rotating the inner cylinder, the liquid in contact with the surface of the rotor acquires its velocity, while the liquid in contact with the wall of the fixed STUDIES IN GELS 91 cylinder remains at rest. As shown by Fig. 64b, this gives rise to a velocity gradient in the gap and thus to a force couple on the rodlets dissolved. This force couple, however, is counteracted by the Brown- ian movement of the particles, which tends to annihilate the orienta- tion brought about by the shear. As a result of this competition 'Z, u \ ^ \: y b) Fig. 64. a) Scattering of rodlets when oriented; a angle of scattering, b) Orientation of rodlets by a gradient of flow. Zj revolving inner cylinder, Zg immobile outer cylinder, d gap between the two cylinders, u maximum velocity of flow, u' velocity' of a rodlet which is oriented by the velocity gradient. between orienting forces and Brownian movement, the rodlets are scattered with respect to the axis of orientation. The distribution function of the rodlets is very complicated, but can be derived from theory (see, for instance, Peterlin and Stuart, 1943)- It is found that the direction of the axis of orientation depends on the length of the rods. With short rods (axis ratio a:b ^ i) the orientations are spread about an axis enclosing an angle of 45° to the direction of flow. With increasing length of the rods (a:b> i), the axis tends to be oriented in the direction of flow, finally (when a : b -> 00) becoming parallel to the tangent plane of the cylinder. The direction of the axis of orientation can be ascertained in the polarization microscope by the direction of extinction. The extinction angle therefore provides in- formation as to the length of the micelle rodlets or macromolecules 02 FUNDAMENTALS OF SUBMICROSCOPI C MORPHOLOGY I dissolved, since short particles give extinction angles of about 45°, whereas filaments give angles near 0°. Having determined the extinction angle, one can also measure the retardation (technical notes in Wissler, 1940, and historical review in PiLNiK, 1946). The birefringence of flow is not a constant as is the double refraction of crystals, because the retardation does not only depend on the thick- ness of the layer, but also on the velocity gradient and the viscosity, and on the concentration of the solution. All these variable quantities are combined in Maxwell's constant, by which the anisotropy of flow of different sols can be characterized and compared. With sols in which the particles of the solute are chain molecules (molecular colloids), the method can be used to obtain data on the anisotropy of single macro molecules. In the case of single chain molecules we can no longer speak of refractive indices, since the surface of a molecule does not represent a phase boundary where the velocity of propagation of light is changed by a definite amount. The optical properties of the molecules are therefore characterized by another quantity, designated as optical polan':(ability, -which, is a measure for the influence of the electromagnet- ic field of a fight wave on the orbits and oscillations of the electrons in the molecule. This influence depends on the direction of vibration of the fight, and in a rod-shaped molecule with rotational symmetry we must therefore distinguish two different principal polarizabifities, the one parallel and the other perpendicular to the mohcule axis, in the same way as we must distinguish two principal refractive indices in an optically uniaxial crystal. More than once the question has arisen (e.g., Schmidt, 1938) as to whether chain molecules, like micellar strands, cause rodlet bire- fringence when they are in parallel alignment. This problem has been solved by Sadron (1957). It follows from the theory developed by him that the formula for the double refraction of flow consists of two parts. The first part depends only on the polarizabifity of the molecule (compare intrinsic double refraction), whereas the second part contains also the influence of the shape of the particles (compare form bire- fringence). In contrast to the conditions prevailing in micellar systems, however, both terms depend on the refractive index of the solvent (Snellman and BjornstAhl, 1941). 4 STUDIES IN GELS 93 The birefringence of flow has furnished arguments in favour of the view that the micellar strands of protein gels are beaded chains (Fig. 51a, p. 66). A flowing solution of 1.5% gelatin is isotropic at 40° C. This is an indication that this sol contains globular protein molecules. When cooled down to 20° C. the gelatin sets after some time. During the incipient gelification the solution becomes birefringent owing to the formation of micellar strands. The extinction angle of the double refraction of flow permits calculation of the length of the elongated particles in such a gel solution. Whereas a diameter of only about 50 A must be attributed to the globular protein molecules, the measured chain length is more than 1000 A and it increases steadily up to over 6000 A before the system solidifies. Particles of this length could not possibly be formed by unfolding of the polypeptide chain, which is somehow coiled in globular protein molecules ; its cross-section measuring about 46 (A)^ (see p. 365), its length could not exceed 1500 A when coiled in a sphere of 50 A diameter. Joly (1949) therefore concludes that the micellar strands result from linear aggregation of globular macromolecules forming beaded chains. When these have become sufficiently long, they interact and a three dimensional network, i.e. a gel, is formed. This gel, containing 1.5% gelatin by volume and micellar strands of 50 A diameter, must have a relatively wide-meshed network. Assuming that the beaded chains meet with the tetrahedron angle of 109.5°, the edges of the polyhedra which compose the framework are as much as o.i /x long. The force of aggregation in these gelatin chains is weak. By increasing the velocity gradient in the apparatus inducing birefringence of flow, the micellar chains of a gel solution of gelatin are shortened by rupture. The applied force couple is of the order of Van der Waals forces, an indication that no valency bonds have been formed between the beads of the chain. This is the reason why a gelatin gel can be melted and reduced to a sol by simple heating. According to Joly (1949), the same beaded chains are formed when proteins with globular molecules are denatured (seep. 136), e.g. when a solution of blood albumin is heated. At a certain temperature intra- molecular bonds are loosened and become free to replace the Van der Waals bonds between the aggregated molecules by chemical bonds, such as hydrogen-, salt- or ester-bonds (see p. 145). Then the protein has become insoluble and, therefore, denatured. Similar observations have been reported of ovalbumin by Foster and Samsa (1950). This protein consists of relatively small globules (Fig. 2, p. 11) which can be unfolded by a high flow gradient to sinuous chains of 600 A length. But this occurs only when the con- centrations are low (< 0.6%). Particles of 2000 A length have been measured in more highly concentrated solutions (2.4%). Such lengths 94 FUNDAMENTALS OF SUBM IC RO S C OP I C MORPHOLOGY are only possible if several particles aggregate. It is unlikely, however, that the aggregation affects fully folded globular particles; probably they become deformed and partly unfolded by the flow gradient, so that somewhat expanded macromolecules aggregate. Micellar textures. Some examples will demonstrate the results ob- tained so far in the optical structure analysis of gels (Frey- Wyssling, 1930). The majority of gels to be considered possess a micellar frame- work containing regions of lattice order with rod-shaped crystals. In the following schemes these are indicated by dashes, although it should be remembered that these lattice regions do not represent isolated dispersed particles but that they are all interlinked and inter- woven by chain molecules. When it has been ascertained by a combination of optical results and X-ray analysis, or birefringence of flow, that the rod-shaped lattice regions or the chain molecules are optically positive with respect to the longitudinal axis, the orientation of the lattice regions can be derived from the character of the double refraction in various sections of the gel. This can be demonstrated in particular in the case of all walls ot anisodiametric plant cells. As shown in Fig. 65, the orientation of the lattice regions is indicated by the arrangement of index ellipsoids in radial, tangential and cross-section. In the secondary wall of a bast fibre the lattice regions run almost parallel to the axis {fibre texture. Fig. 65a). If their orientations are scattered with respect to the cell axis, the cross-section which in the first case is almost isotropic becomes birefringent ; we obtain a fibroid texture (Fig. 65b). The counterpart of the fibre texture is the ring texture (Fig. 65c), in which all lattice regions run in tangential orienta- tion. This texture occurs in the ring-shaped reinforcements of young vascular cells. If, starting from this texture which is optically negative with respect to the cell axis, the lattice regions are allowed to scatter, the widespread tube texture is obtained (sieve tubes, latex tubes, vessels, elongated parenchyma cells, etc.). Here the tangential section is optic- ally negative; the radial section, however, is positive, since all projec- tions of the scattered rod-shaped lattice regions upon the radial section are approximately parallel to the axis. As there is a continuous change from the negative region to the positive one, a front view of these cells will show an isotropic zone in which the two regions of opposite sign become merged (Fig. 65d). STUDIES IN GELS 95 If the lattice regions do not scatter, but deviate from the direction of the cell axis while remaining parallel to each other, a spiral texture is obtained, as occurs in cotton wool fibres, the tracheids of conifers (Jaccard and Frey, 1928; Preston, 1934, 1946) and the wood fibres of deciduous trees. H-'^ -j-""? riy ny o) + ii/mi nil c| l|l, lllll« l/lIU llllll lllllll llllll llllll lllllll l''l|l| IHllU ll'll'l llllll II Ml Ill/Ill III' ll>llil + + '"'lllllll // 'INN ■A « i/,i /i\ i;:('* /\<\ii llllll 'llill' V ,7) \i 11111/ llllll lllllll !ll| lllllll llllll mm nil iilVii lllllll mil lllllll iiiii mini 'mil 1 mil nil; Hy fly I — I I - I .. *^— ^- __'ri ■ ^'~ ~~ I't'i rv\ ;','•' ,' 1. '.'• ', ■ ^— ^ .,''1 <'- ii'''i •:-■■: ^ zz. ^ •'••'/• V'l' ,tl .< •Ill, 'IJ, l'(l'.' r: I .1' / [I 1 ''1 ||.;. I'l" 1 "1 5<; :i I'l" 'ill 'll,' iii' !'' •ii' ii';; \ ll '-^ *.' .' "Z-^ ^ 'J B": c) ^.V-i-'n'y d) Fig. 65. Micellar textures of cell walls (from Frey-Wyssling, 1930). a) Fibre texture, b) fibroid texture, c) ring texture, d) tube texture, n^ biggest, Hq smallest refractive index of cellulose; n^* biggest, n^* medium, n^* smallest refractive index of the cell wall, i isotropic, -f optically positive, — optically negative. In isodiametric objects there exists no morphological axis which may serve as reference axis to the double refraction. In spherical objects such as starch grains, spherites and the like, the radial direction is therefore chosen as reference axis. If the refractive power for vibrations parallel to this axis is larger than that for vibrations in a tangential direction, the spherite texture is called positive; in the opposite case it is called negative. The determination of the optical character of a spherite built up of chain molecules or rod-shaped lattice regions, however, does not sufhce to derive its submicroscopic texture. For, 96 FUNDAMENTALS OF S UBM I C RO SC OP I C MORPHOLOGY I as shown in Fig. 66, spherites can be positive or negative both with radial and with tangential arrangement of the structural elements, depending on whether the structural elements themselves are positive or negative with respect to their axis. Hence the first thing to ascertain is the optical character of these structural elements. In most cases the texture is as shown in Fig. 66a (starch grains, inulin). Fig. 66. Gels with spherite texture, a) Positive, b) negative spherite of positive rodlets; c) negative, d) positive spherite of negative rodlets; e) positive myelin sphere (oblate) of positive rodlets. In hollow spheres, the reference axis cannot be determined with certainty. In isodiametric parenchyma cells, for instance, the double refraction of the cell wall is referred to the tangential direction (Fig. 63e, p. 87), in analogy to the situation in anisodiametric cells, although they are isotropic in radial direction. This is due to random orientation of the structural elements in the tangent plane. An arrange- ment of this kind is designated as foliate texture. For further details of optical texture analysis we must refer to the literature concerned (Ambronn and Frey, 1926; Frey, 1926b; Frey-Wyssling, 1930, 1935a; Schmidt 1934, 1937a). b. X-ray Analysis of Gels Micellar strands. A complete structural analysis by means of X-rays is only possible if crystalline lattice regions are present. In the case of molecular colloids such as rubber solutions, protein solutions, etc., irradiation with monochromatic X-rays furnishes as a rule no more than an "amorphous" ring, which gives some information about the /«/ramolecular periods occurring most frequently (for instance, in rubber: the length of an isoprene unit). Only when the chain molecules are arranged in a crystal lattice does X-ray analysis yield interference phenomena rich in lines or spots, from which far-reaching morpholog- ical conclusions can be derived. This will be further illustrated bv 4 STUDIES IN GELS 97 means of the cellulose diagram of ramie fibre (Fig. 67), Each point on the diagram corresponds to a set of parallel net-planes in the crystal lattice. The diagram of Fig. 67 enables us to measure four quantities : I. the mutual distances, 2. the density, 3. the breadth and 4. the arrangement of interferences, each of which permits calculation of a corresponding quantity in the undisturbed lattice regions. Fig. 67. X-ray fibre diagram from ramie showing lavcr lines. I. According to Bragg's law of reflexion, the distance between the lattice planes is calculated from the distance between the interferences and the centre of the diagram. We learn from X-ray optics how the unit cell (see p. 26) in the crystal lattice of cellulose can be computed from the distances measured in the diagram of artificially oriented cellulose preparations whose crystalline regions display an arrangement of even higher orientation than in ramie fibres. The elementary cell found for crystalline cellulose is monoclinic; its sides are a = 8.35 A, b = 10.3 A, c = 7.9 A, and the angle ^ between a and c is 84° (Meyer and Mark, 1930). Of these quantities, the most accurately determined is the fibre period b which corresponds to the length of a cellobiose molecule (Fig. 68). It is calculated from the distances between the so-called layer Hnes which are clearly visible in Fig. 67, running parallel to the equator of the diagram and connecting, as it were, the inter- ference spots. These interference spots are broadened along the layer q8 fundamentals of submicroscopic morphology I lines as a result of cellulose chains which do not belong to the crystal lattice (Sauter, 1937). 2. From the density of interferences the number of atoms in the net- planes can be derived, since the lattice planes reflect the X-rays more intensely in proportion as they contain more atoms. The density of interferences can be estimated, or measured photometrically. In Fig. 67 two black spots can be seen on the equator, with a mutual distance of 261/4 i^rn. Their great density is caused by the family of net-planes which contain the glucose rings of the cellulose chains and, as both points correspond to the front plane of Fig. 68, the ring of the glucose units must lie in this plane. In this way it is possible from the intensity of the interferences to determine the orientation of the molecular models (obtained on structural chemical grounds) in the unit cell (derived from X-ray analysis). 3. From the breadth of the interferences one can calculate the width of the undisturbed lattice regions, using a method developed by Scherrer (1920) for metals, i.e., substances absorbing X-rays, and worked out by Laue (1926) for non-absorbing substances. To do this the density must be measured photometrically. The breadth at half- maximum of the density peaks in the photometer curve (Fig. 70, p. 102) is a measure of the dimension of the crystalline regions perpendicular to the set of net-planes causing the interference. The broader the X- ray interference in the diagram is, the smaller is this dimension. In Fig. 67 the interference spots on the equator are clearly broader than those near the poles of the diagram. It follows from this that, in the fibre, the dimensions of the lattice regions are considerably smaller in directions perpendicular to the fibre axis than in directions parallel to this axis. They must, therefore, be rod-shaped, in conformity with the conclusion drawn from the character of their form birefringence. Hengstenberg and Mark (1928) find 50-60 A for the thickness of these rodlets. Their length cannot be measured accurately, the for- Fig 68. Crystal lattice of cellulose (from Meyer and Misch, 1937). 4 STUDIES IN GELS 99 mulae being very insensitive to changes in lengtii when this length is large (Frey-Wyssling, 1937a). The experiments admit of no doubt, however, that the length of the rodlets must be more - probably much more - than 10 times as long as their thickness. 4. From the arrangement of interferences can be derived the arrangement of the rod-shaped lattice regions. In the diagram considered all rodlets are parallel to the fibre axis {fibre diagram) ; but if they follow a screw line within the wall, the interferences on the equator are drawn out into sickles {.uckle diagram). Finally, if they lack all order, interference rings instead of spots are obtained (Debye-Scherrer or ring diagram^ see Fig. 69, p. 100). A comprehensive and simple treatment of the relation between the arrangement of interferences and that of lattice regions has been given elsewhere (Frey-Wyssling, 1935a, p. 11). Ring, sickle and fibre diagrams are represented in Fig. 75, p. 106). Working out the fibre diagrams in full detail from the four points of view mentioned, one arrives at the structural model of the fibre wall shown in Fig, 59b (p. 77). This picture renders all the facts which can be ascertained by means of X-rays, though the rodlets are in reality much thinner. When drawing such a scheme it should always be borne in mind that X-ray analysis only gives information about the regions of lattice order; no information can be obtained in this way about the regions without lattice structure. In particular, it cannot be decided by means of X-rays whether the chain molecules in the crystal lattice are of exactly the same length as the lattice regions or whether (as has already been mentioned) they protrude from these regions without order and invade several other lattice regions (Fig. 54, p. 70). X-ray analysis therefore tells us nothing about the manner in which the crystalline regions are interlinked or about the interstices between the regions of lattice order. From a biological point of view, however, these intermicellar spaces are of special importance. For, in most substances possessing a frame- work, the micellar strands with their crystalline regions are to be con- sidered as virtually lifeless, while all perceptible processes of life presumably take place in the intermicellar system. Thus, the mechanical properties of a gel are determined in the first place by the micellar structure, whereas for all physiological questions (such as permeability, metabolic processes, vital staining, etc.) one should study primarily the intermicellar regions. lOO FUNDAMENTALS OF SUBMI C RO SC OPI C MORPHOLOGY I Intermicellar spaces. The regions between the meshes of the micellar framework may represent a homogeneous phase if they are filled with a uniform liquid or gas. This only holds good so long as the gel frame consists of strands which can themselves be considered as a phase, so that a phase boundary exists. If the strands become so thin, however, as to reach the dimensions of a chain with the thickness of a single molecule, the concept phase loses its significance. (ni) (002) (022) (113) (222) Fig. 69. X-ray diffraction pattern of a) ramie and h) %\W stained with gold. In addition to the fibre diagram, Debye-Scherrer rings of gold (m), (002) etc. are seen (from Frey- Wyssling, 1937a). Information as to the dimensions of the intermicellar spaces in the gel frame can be obtained in various ways. If one succeeds in filtering particles of known size through a gel, the inference is that the pores are bigger than the particles, as in an ultrafilter. Unfortunately, how- ever, it is not possible to obtain absolute values of the pore size of the intermicellar spaces with the aid of ultrafiltration (Czaja, 1950), since differences in electric charge or in chemical behaviour (hydrophoby) very strongly influence the ease with which filtration of the particles takes place (Morton, 1935). For this reason, only relative sizes can be obtained, which cannot be compared with the absolute values deter- mined by means of X-rays. Until now it has not been possible to obtain X-ray diagrams of the intermicellar substances ; for, even when in the solid state, they do not usually show the properties of crystals. In the plant cell wall, e g.. 4 STUDIES IN GELS lOI silicic acid, lignin, etc. are embedded in the amorphous state and there- fore do not produce X-ray interferences. For this reason the amount of space occupied by the intermicellar regions in frameworks was totally unknown. To obtain information in this important field, foreign substances must be introduced into these spaces, where they crystallize and can then be submitted to X-ray analysis (Frey-Wyss- LiNG, 1937a). We must therefore create by artificial means an inter- micellar substance possessing lattice order, which enables us to derive quantitative data of the dimensions of the unknown submicroscopic regions. Gold and silver crystals have proved to be the most suitable for this purpose. Following Ambronn, the objects are soaked in 1-2",', solutions of gold chloride or silver nitrate, then carefully dried with blotting paper and finally the salt absorbed is reduced by means of light or hydrazin hydrate (Frey, 1925). In this way microscopically homogeneous colourings are obtained displaying a beautiful di- chroism (compare Ambronn and Frey, 1926, coloured table; Wiener 1926a). The X-ray diagram of the dyed fibres shows Debye-Scherrer rings of crystalline silver or gold (Fig. 69) in addition to the fibre diagram of the framework substance (ramie fibre, silk and wool). The annular interferences prove that the metal crystallites imbedded take up all possible positions with respect to the fibre axis. The size of the cubic gold and silver crystals is calculated from the breadth at half- maximum of the interference rings (Fig. 70). The investigation produced the surprising evidence that metal crystallites with a cross-section of about 50 A are incrusted in silk and wool, and particles even exceeding a diameter of 100 A in ramie fibres (Table XII). Since the strands of the micellar framework in ramie fibres have a thickness of only 50 A, the artificially embedded metal crystallites occupy an unexpectedly large space. Notwithstanding their great strength, cellulose fibres must, therefore, be built rather loosely, a fact which was already known from density measurements in the bleached fibres used in these experiments. After removal of all foreign substances, the density of ramie fibres amounts to only 1.39, whereas the density of cellulose is 1.59. There should therefore be about 12.6% of submicroscopic empty space* (Frey-Wyssling and Speich,i942). ^ The density 1.39 ~ 0.03 is derived from accurate determinations of mass and volume. If, instead of the density of crystalline cellulose, one uses the density 1.55 of the incom- pletely crystallized fibre measured in toluene, one finds a discrepancy of 10.3% . 102 FUNDAMENTALS OF SUBM IC ROSC OP I C MORPHOLOGY TABLE XII PARTICLE SIZE A OF GOLD AND SILVER CRYSTALS EMBEDDED IN FIBRES Metal embedded A in A Ramie fibres Ag 85 Ramie fibres Au 84 Hemp fibres Au 90 Bamboo fibres Au 85 Wool Au 58 Silk Au 50 It is clear that not all cellulose rodlets with a cross-section of about 50 A can be surrounded by spaces 100 A wide, as otherwise the dis- crepancy in density would be much greater still. Furthermore, the phenomena of swelling require very narrow intermicellar spaces of the A I •/- Fig. 70. Photometer curve of hemp fibres stained with gold (the distance and breadth ot the interferences are magnified 2.0 times as compared with Fig. 69). From the breadth at half-maximum of the dcnsitj' peaks the diameter yl of the embedded gold crystals can be calculated (from Frey-Wyssling, 1937a). order of 10 A, into which the water can penetrate, pushing the cellulose rodlets apart. In dyed ramie fibres there must therefore be two cate- gories of submicroscopic spaces, viz., i. narrow intermicellar spaces of 4 STUDIES IN GELS 105 the order of magnitude 10 A which are responsible for the phenomena of swelling, and 2. wider capillary spaces which are accessible to dyes of much larger dimensions and to the hardening substances Hgnin, cutin, etc. For this reason they are of primary importance technically in the process of dyeing and physiologically in the hardening of the cell wall. It must be supposed that these larger spaces are widened by the growth of the substances embedded. The capillary shape of the wider spaces can be proved in the follow- ing way: in objects with a well-developed fibre texture the gold and silver particles embedded give rise to a strong rodlet dichroism (Frey-Wyssling and Walchli, 1946). This is only possible if the isodiametric metal crystals are arranged in rows or in rod-shaped aggregates; i.e., the metal particles must lie in pre-formed submicro- scopic canals. Even more can be inferred from experiments with silver amalgam. If mercury is precipitated in the fibre from an alcoholic solution of sublimate, dichroic colouring is obtained which does not produce an X-ray diagram, because the mercury is present in the Uquid Fig. 7 1 . Oriented embedding of silver amalgam in the fibre. In addition to the fibre diagram of cellulose (broad interference spots) a fibre dia- gram of silver amalgam (narrow interference spots)appears (frompREY- Wyssling, 1937a). state. Treating the fibres afterwards with a solution of silver nitrate, one obtains X-ray diagrams showing interferences of silver amalgam (Fig. 71) in addition to the diagram of cellulose. The silver amalgam c.lOO^' 104 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I crystallizes in submicroscopic hexagonal needles which all run parallel to the fibres axis, for, instead of a Debye-Scherrer diagram, 3. fibre diagram of silver amalgam is obtained. This proves the presence of submicroscopic canals in the fibre. It is much more difficult to discover the dimensions of these pre- formed capillaries, as the size of the gold crystals embedded varies with the speed at which they develop in the capil- lary system. Furthermore, the metal rodlets causing the dichroism are so large that they can easily be shown in the ultramicro- scope (Fig. 72). Their rod- shape is betrayed by the different intensity of the light scattered in lateral irradiation according as the vibration of the line- arly polarized light is par- allel or perpendicular to the fibre axis (Frey-Wyss- LiNG, 1937b). Consequent- ly, the crystals or primary particles measured by means of X-rays must have clustered together to form rod-like aggregates or secondary particles (Fig. 73), widening the capillaries in so doing. Whether this already takes place while the crystals are growing, we do not know. Presumably, however, they can develop fairly freely, since they do not acquire the rodlet shape of the capillary system until they are collected in the secondary crystalline particles. We must there- fore take it as proved that, apart from the intermicellar spaces in which the water penetrates when the fibre swells, there exist even larger pre- formed inhomogeneities . As a consequence, native fibres must possess long-shaped submi- croscopic regions containing intermicellar spaces which are only accessible to small molecules such as water, salt ions and iodine. These regions are designated as microfihrils\ they possess a more or less homo- Fig. 72 Fig- 75 Fig. 72. Ramie tibre stained with silver in the ultra- microscope (from Frey-Wyssling, 1937b). Fig. 73. Ultrastructure and shape of the submicro- scopic silver rodlets in the fibre. STUDIES IN GELS 105 capillary structure (Fig. 74a). In between these microfibrils, however, interfibrillar capillaries must occur in the form of wider canals, in which larger molecules such as colloid dyes and incrusting material are de- posited. The porous system of ramie fibres is, therefore, heterocaplllarj ; the smaller intermicellar spaces (of the order of 10 A) and the larger interfibrillar ones (of the order of 100 A) communicate freely. The 1000^ = 0.!fi i(ca 10 A } mfca 60 J) Ifca 100 ^j b) o) Fig. 74. Miccllar structure of bast tibres (from Frey-Wyssling, 1936a, 1937a). a) Longitudinal and cross-section; intermicellar spaces all similar (homocapillarity).Z') Cross- section with coarser interfibrillar and finer intermicellar spaces (heterocapillarity); microfibrils f composed of micellar strands m ; i intermicellar spaces, k interfibrillar capillaries. microscopically visible fibrils must still contain both categories of spaces, because. as a rule they can be dyed like the whole fibre, and they thus represent aggregate bundles of the invisible submicroscopic microfibrils. Stretching experiments. A subject which has become of special im- portance in the study of gel structure is the X-ray analysis of the/)ro^(?j-j- of orientation in stretching experiments. By way of example we shall briefly go into the phenomena observed in stretching regenerated cellulose fibres obtained from viscose. It is possible to make isotropic cellulose fibres from viscose (Hermans and De Leeuw, 1937). The X-ray diagram of these fibres Io6 FUNDAMENTALS OF S U B.M I C RO S COPI C MORPHOLOGY I consists of Debye-Scherrer rings. If the orientation of the micellar strands is completely random, photometric measurements show the intensity round each ring to be constant. If, now, the isotropic fibres are stretched, the micellar strands are oriented. With increasing stretch, the X-ray diagram changes into a sickle diagram and finally into a fibre diagram (Fig. 67, p. 97) when orientation is complete. If, at a given (1 ) hO 1.25 1.62 f» .| 1.89 Fig. 75. X-ray diagram of Hermans's threads, gradually stretched. The numbers give the degree of stretching (length of stretched gel/original length). (From Kratky, 1940). degree of stretch, one measures the intensity along the interference sickles corresponding to the equator interferences in the fibre diagram (paratropic interferences), the/r^^//^;z(7 with which the different orienta- tions of the micellar strands occur can be derived from the decline in intensity from the equator towards the poles. In fact, the intensity depends on the number of lattice planes which take part in the reflexion of X-rays. It is possible in this way to determine experimentally the distribution function of the orientations of micellar axes. 4 STUDIES IN GELS IO7 If the distribution were one which covers a sector with uniform density (Fig. 64a, p. 91), as was assumed on p. 90, the sickle inter- ferences would be circular arcs with sharp boundaries, extending over a sector angle dependent on the angle of scattering. As shown by Fig. 75, however, the density in the sickle decreases very gradually towards the poles, and the distribution function is a very complicated one : the micellar strands which enclose a small angle with the direc- tion of the stretch are more frequent than 'those which form a large angle with this direction, and this distribution is a function of the degree of deformation (Hermans, Kratky and Treer, 1941). In order to explain the distribution curves found experimentally (in- tensity depending on angular distance from the equator), and their change with the degree of stretch, Kratky (1940) has made two different assumptions with regard to gel structure and has calculated how the distribution alters in the stretching process. Comparing these theoretical curv^es with those obtained experimentally, it is possible to decide which of the two hypotheses is the more likely. The first limiting case considered by Kratky (1935, 1940) conforms to the older ideas about gel structure, assuming rod-shaped "freely suspended micelles", which are independent of each other (Fig. 59b, p. 77). Their orientation in the stretching process is achieved, as it were, by the flow of liquid (swelling medium) which turns the rodlets distributed at random into positions which are parallel to the direction of the stretch. On this assumption the distribution of the micellar orientations can be calculated for any degree of stretch (= final length divided by original length of the gel). Advanced parallel arrangement of the rodlets is only reached at high degrees of stretch. A number of very swollen gels of cellulose esters (cellulose amyl oxalate, trinitro- cellulose) show a behaviour which is in conformity with this theoret- ical distribution. On the other hand, it seemed surprising at first that, in the case of relatively low degrees of swelling (between 1.5 and 2), neighbouring micelles do not disturb each other's movements and behave according to formulae which have been derived for particles freely suspended in a large amount of liquid. To explain this, Kratky (1954) suggested that the arrangement of micellar rods is not completely random, but that there must exist short-range order (i.e., order in small regions). This means that if small, submicroscopic regions are considered, a certain I08 FUNDAMENTALS OF SUBMI C RO SCOP I C MORPHOLOGY I parallel arrangement is founds At some distance, however, the ar- rangement becomes gradually more and more disturbed, so that all possible orientations are found in a gel volume of even microscopic dimensions. This is shown by Fig. 76. Hence, when considering the dispersion of orientations in Fig. 51b (p. 66) or 54a (p. 70), it must be borne in mind that neighbouring particles are almost parallel. An entirely dif- ferent orientation is only found at a certain submicroscopic distance as a result of grad- ual changes in orientation. In the much larger microscopic dimensions this means. >>\nOa\\ i/i,n\l Fig. 76. Short-range order of however, that all anisotropy effects are neu- short rod molecules (from ^ralized as if a random criss-cross arrange- Hermans, 1941). . ment existed. The principle of short-range order would explain why it is that, when stretched, gels of a low degree of swelling can behave as if their particles were freely floating micelles. In fact, the movement of each particle is very similar to its neighbour: there is no steric hindrance, as would be the case if the arrangement were an irregular one. The principle of short-range order does not suffice, however, to explain altogether the behaviour of gels when stretched ; for, the extensibility of these gels would have to be unlimited, and it should be possible to deform them to fibres of arbitrary length, even in those cases where the degree of swelling is low. In the cellulose fibres mentioned, prepared by Hermans, this is im- possible. We are therefore forced to assume that the micelles are not freely movable, but that they are interlinked by junctions (Frey- Wyssling 1936a, 1936c) or hinges (Fig. 77). This assumption of complete interlinking of the structural elements in the gel is designated by Kratky as the second limiting case. Here again, there exists short- range order, and the picture arrived at (Hermans, 1941) corresponds more or less to the one given by us (compare Fig. 54, p. 70). In other words, the orientation takes place as if chains consisting of rigid links and movable but inextensible hinges were stretched by pulling at the ^ The voluntar\'^ parallel arrangement of rod-shaped particles is not confined to colloid raatter.lt occurs also in pure liquids and real solutions, where physicists speak of short- range order (Zernike, 1939; Stuart, 1941; Peterlin and Stuart, 1943). Taking an arbitrary molecule, its immediate neighbours are more or less orderly as regards distance and orientation. STUDIES IN GELS 109 ends. On this assumption a distribution function for the orientation in network systems can be derived. A striking result of this theory is, that a completely parallel arrangement of all micellar strands would be reached at a degree of stretch 2 (100% stretch). This is not in keeping with the observed facts, seeing that Hermans's cellulose fibres, especially when greatlv swollen, can undergo a stretch of several times 100",,. One must, therefore, assume that in reality neither the first nor the second limiting case is realized. The behaviour is intermediate between those corresponding to the two extreme cases calculated, i.e., the micelles are not freely suspended but they are inter- linked to form micellar strands. The junctions present, however, are not fixed indis soluble hinges which completely prevent the micellar strands or parts of these strings from gliding past each other. In fact, the cohesion must be due to forces which at certain points can be overcome by the orientating forces, so that a "flow in small regions" takes place. c. Swelling of Gels If isotropic gels are immersed in a swelling medium, they swell uniformly in all directions. If a certain orientation of the micellar strands prevails, however, the swelling is anisotropic, i.e., different in different directions. The anisotropy of swelling of starch grains induced Nageli (1858) to consider the structural units (micelles) of the gel as submicroscopic rodlets. Intermicellar swelling. According to Nageli, the swelling medium penetrates between the rodlets which we now call the micellar strands. In many cases X-ray analysis has confirmed this view, as often the X-ray diagram does not change in the swelling process, so that ap- parently the crystalline regions remain unaltered (e.g., plant cell walls and cellulose gels). Fig. 77. Short-range order in a gel of inter- linked micelles (from Hermans, 1941). no FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I The swelling medium penetrating between the string4ike structural elements causes the system to inflate laterally. For this reason swelling is always at its greatest in directions perpendicular to the direction of orientation of the micellar texture, and is almost zero along the fibre axis if the fibre texture is ideal. The arrangement of the micellar strands can therefore be derived from the anisotropy of swelling, or conversely, the anisotropy of swelling or shrinking to be expected can be computed from the optical anisotropy measured (Steinbrinck, 1906; Ziegenspeck, 1938). If it is assumed that the microfibrils of native fibres, made up of polyhedral micellar strands, possess a more or less circular cross- section (Fig. 74b, p. 105), these can be idealized as circular cylinders. It is then found that in the completely dry state 9.5 per cent, by vol. of intermicellar empty spaces must occur between the strands (Hermans, 1938). This value tallies approximately with the average empty space (8.5 %) obtained from determinations of double refraction and density (Frey-Wyssling and Speich, 1942), showing that in well-dried fibres the microfibrils are fairly closely packed. Gels in which the colloid portion is crystalHzed imperfectly, so that a large amount of amorphous substance is present, swell much more than well-crystallized fibres, the swelling medium being able to penetrate into the unordered regions, causing them to swell. Nevertheless it does not succeed in solvating the individual chain molecules in the ordered regions. Intramicellar swelling. If, however, the affinity between the swelling medium and the chain molecules is stronger than the binding forces in the chain lattice, the swelling medium will penetrate into the lattice and widen it. This widening can be followed by means of X-rays and is often found to abolish the interferences. In that case the chain molecules are completely solvated and if they are not kept together by valence bridges (p. 67), unlimited swelling can take place which will gradually lead to the dissolved state of a sol. In many cases, however, swelling media react with the side groups of the macromolecules, causing a change in the chemical character of the high polymer chains. This applies, for instance, to the esterifica- tion of solid cellulose (nitration, acetylation, Frey-Wyssling, i936d). If the changed chain molecules cannot be solvated by the penetrating swelling medium, the result is a lattice of the newly formed substance and no unlimited widening of the chain lattice takes place. This 4 STUDIES IN GELS III phenomenon, too, can be followed by means of X-rays, since the new chain lattice shows new interferences, while the orieinal ones disappear. These conversions are termed permufoid or topochemical reactions, because the reacting groups undergo chemical changes within the crystal lattice itself without dissolution of the molecules. The characteristic feature of these reactions lies, therefore, in the fact that chemical changes take place in the solid state, in contrast to the classical formula: corpora non agiint nisi fluida. Intramicellar swelling clearly demonstrates the great similarit}- between swelling and dissolution. As has been shown by Katz (1924), in both cases the same physico-chemical phenomena take place (heat of swelUng, volume contraction and swelling pressure as a result of solvation), the only difference being that swelling occurs very slowly because of the slow Brownian movement of the macromolecules. And if in some way or other these form a network, only limited swelling takes place and the state of a sol is not reached. Shrinkage. Most gels encountered in nature are liable to swell to a certain extent. On drying, the behaviour depends on the properties of their gel frame. If this possesses meshes with fixed contours, such as, for instance, silica gels, the decrease in volume does not correspond tothelossof water. The dry system is a porous body, i.e., it has changed into an air-containing aerogel. If the gel framework is flexible, however, the meshes will graduallv close on continued shrinkage till finally the micellar strands touch on all sides. The result is a horny, brittle xerogel without perceptible porosity. The drying process of these xerogels is very problematic. If we assume the gel to be isotropic, it must possess a gel frame of random arrangement. Were we to apply this principle of randomness also to amicroscopic regions (Fig. 53a, p. 69), the framework obtained when the molecular or micellar strands approach each other would be a loose structure with numerous interstitial or intermicellar spaces. In that case the xerogel would possess a lower density than the crystalline substance and it would have a white and untransparent appearance as a result of the light diffraction caused by the air-containing spaces. This only applies, however, to aerogels, whereas xerogels solidify to completely transparent glassy substances. If the density of the crys- talline micellar strands is determined by mean? of X-rays and compared with the density of xerogels, the discrepancy found is only about 10% 112 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I (Hermans, 1938), whereas a dried mass of micellar strands should represent a more airy structure with a much lower density. Examples of xerogels are gelatin and celloidin. We are therefore compelled to assume the existence of short-range order. Given this short-range order of the micellar strands, one can imagine continuous strings intersecting the whole gel (Fig. 54, p. 70). The orientation never changes abruptly; deviations from parallel alignment are only gradual. Following such a continuous string or micellar strand in an isotropic gel, one finds a curve; neighbouring strands are approximately parallel (Fig. 85, p. 127). Shrinkage causes the strings to approach each other; if the distance between them re- mains the same at all points, the result must be a decrease in the radius of curvature (Hermans, 1941). It follows that, on the assumption of short-range order, the gel is capable of shrinking uniformly in all directions until the structural elements are close-packed, without kinks in the micellar strands (Fig. 76, p. 108). Discrepancy in the density of dry gels. The transparent brittle state of dry xerogels (dried glue, gelatin foil, horny celloidin, etc.) has led Hermans and Vermaas (1946) to compare these substances with glass. In the manufacture of glass the rapid cooling of melts gives the un- wieldy molecules of quartz, silicates, borates, etc. no time to crystallize. The glassy amorphous state is, therefore, characterized by a similar molecular framework to that of the gels with amicroscopic framework, i.e., with chain molecules as structural units. Glasses possess a some- what lower density than crystals of the same compound, since the closest packing of the molecules is attained in the crystal lattice only. For instance, the difference between the densities of butyl alcohol CH3CH2CH2CH2OH in the crystalline and in the supercooled glassy state amounts to 6%. Gels with micellar structure contain ordered crystalline regions of micellar strands next to less ordered, more or less amorphous regions. For the latter Hermans (1946) assumes an amorphous glassy state. Hence the gel consists of crystalline and glassy amorphous parts. If the densities of the crystalline and the amorphous compounds are known, the amount of crystalline material in the gel can be calculated. Using the reciprocal densities, i.e., the specific volumes, the following holds good : X 9?-^^., + (i — x) y^^^^j = 9?, where ff = experimentally determined specific volume of the gel, ^^j^r) = specific volume of the crystalline part, rp^^^^^ = average spec.vol. 4 STUDIES IN GELS II3 of the amorphous part, x = the fraction of crystalline material. Substituting 1.55 for the density of the cellulose fibre (determined in toluene), 1.59 for that of crystalline cellulose and 6% less for amorphous cellulose (compare butyl alcohol), Hermans (1946) cal- culated X = 0.61 for ramie fibres and 0.18 — 0.32 for regenerated cellulose. In other words, only 1/5 to 1/5 of the cellulose in rayon fibres is crystalline. Whereas this result is quite acceptable, the amount of crystalline cellulose in ramie is likely to be greater than 60%. Otherwise the difference between the birefringence of ramie fibres and that of crystalline cellulose ought to be greater than actually determined (Frey-Wyssling and Speich, 1942; according to our measurements it amounts to 4.4% and, based on the double refraction 0.0705 of crystalline cellulose determined by Hermans 1949, to 7.3 %). The crystallinity of a gel can also be determined by X-rays. Since amorphous substances scatter the X-ray beam, they cause a diffuse back- ground blackening of the film in the X-ray camera. The photometer curve (cf. Fig. 70, p. 102) taken from such films permits computation of the amount of the amorphous fraction in the gel under investigation. By this method Hermans and Weidinger (1949) find 70% crystalline cellulose in ramie and 39 % in regenerated cellulose. There is a third means of estimating the amount of the two fractions. As the hydrolysis velocity of amorphous cellulose is much greater than that of crystallized cellulose, the quantitative relation between them can be derived from a suitable hydrolysis/time curve. Philipp, Nelson and Ziifle (1947) calculate by this method 95% crystal- linity for ramie fibres and about 70 % for rayon. As the three methods mentioned (optical. X-ray and chemical) yield different values for the crystallinity of the same gel, we must conclude that there is no net difference between crystallized and amorphous cellulose; hence the non-crystallized fraction is rather to be considered 2lS paracrystalline (cf. Fig. 54, p. 70). Hermans criticizes the opinion that dry xerogels are porous bodies on the ground that no one speaks of submicroscopic spaces in the case of glasses either, notwithstanding the lower density than in the crystal- line state. This comparison, however, does not seem quite justified to me, since certain liquids (such as water, alcohols and aldehydes in the case of cellulose) are capable of penetrating into xerogels, whereas this does not occur in glasses. Thus, clearly, there must exist a differ- ence in the order of magnitude of the "empty spaces" present. In the swollen state xerogels definitely possess a loose structure, and it is not likely that the micellar framework loses this structure completely 0.80 0.70 \j}' r^ "" empty space ^^o' A^^ Water 114 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY Z Upon drying. This is more likely to occur in molecular frameworks. Here the empty spaces shrink and form interstitial spaces which no. longer possess the character of submicroscopic pores. It is therefore easy to see why, in the poorly ^^^^ crystallized rayon fibres, por- | osity will disappear to a § great extent in the drying | process ; all the same, even c^o.90 these fibres contain about 5 % of empty space (Fig. 78). In native fibres whose incrusta- tions have been removed, a complete closing of the struc- ture would hardly be poss- ible. Otherwise it would be hard to explain how the dens- ity of ramie determined in toluene could amount to i . 5 5 , whereas the result of accurate measurements of mass and volume gives only 1.39. Fur- thermore, an inner reserve of space is necessary to explain the great flexibility and capa- city to twist; otherwise these fibres would be as brittle and elastic as glass fibres. Do^ible refraction of swollen gels. In the swelling process, isotropic imbibition liquid penetrates between the anisotropic micellar strands. In this way the rodlet birefringence of gels is enhanced, for it follows from the formula given on p. 84 that if the other conditions remain constant, this birefringence acquires its maximum value when the relative volumes of rodlets and imbibition medium are equal {b^ = ^3)- The intrinsic double refraction, however, is inversely proportional to the volume so long as it is permissible to assume that no change in micellar orientation occurs as a result of swelling. If the intrinsic double refraction of the dry gel is called i-Do and the degree of swell- ing is q (volume of swollen gel/volume of dry gel), then, according; %r 0.60 ' ' "" Cellulose _i • -• " ■ '- 0.!0 0.20 0.30 Regain g H2O per g cellulose Fig. 78. Increase in volume of swelling isotropic (regenerated) cellulose threads (from Hermans^ 1946). Abscissa: absorption of water. Ordinate: specific volume (i/density), q>o specific volume of dried threads 0.66, ^^cr specific volume of crys- tallized cellulose 0.63. The water absorption^ increases linearly, but the volume does not. 4 STUDIES IN GELS TI5 to Kratky and Platzek (1958), the total double refraction of the swollen gel t-Do amounts to : t-Do = t-Do. q Consequently, if the intrinsic double refraction i-Do of the dry gel is known and the total double refraction t-Do of the swollen gel is measured, the rodlet birefringence r-Do of the swollen gel can be calculated. It is therefore possible to measure rodlet anisotropy in gels capable of swelling, provided the birefringence curves are only plotted from points which result from measurements in imbibition liquids giving rise to similar degrees of swelling. Otherwise one would obtain compHcated kinky curves devoid of regularity, instead of smooth Wiener curves (cf. Fig. 62, p. 84). Apart from rodlet double refraction, another form of birefringence mav occur when liquids penetrate between the amorphous chain molecules. This is attributed by Vermaas (1941, 1942) to oriented adsorption of the penetrating molecules. It might also be due, how- ever, to a change in the "intrinsic anisotropy" of the chain molecules caused by the swelling medium, such as that occurring in sols when the refractive index of the dispersing medium is changed (Sadron, 1937)- d. Electron Microscopy Electron rays. The electrons which are emitted by a cathode are electrically charged negative particles with a mass of 1/1840 of that of a hydrogen atom. The range of these electrons in air is very short, because they are absorbed or scattered by atoms or molecules which they meet on their path. All investigations with electron rays must therefore be carried out in vacuo. On account of their electric charge they can be made to deviate from their straight trajectory by means of electric or magnetic fields. Bundles of electron rays can therefore be focused by electric coils in much the same way as light rays by lenses. This makes it possible to form images with electron rays according to the laws of geometrical optics (Zworykin and coll., 1945; Burton and Kohl, 1946; Wyckoff, 1949; Frey-Wyssling, 195 i). In so far as the electron rays represent a stream of particles, they can hardly be compared with light rays or X-rays. They have the remark- Il6 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I able property, however, of possessing at the same time the character of waves. They can be deflected by crystal lattices and, like X-rays, give rise to interferences. Hence an electron ray represents a corpus- cular ray and a wave train at the same time! The wavelength X of electron rays depends on the voltage applied to the cathode tube; A is inversely proportional to the square root of the voltage. In the case of light waves, the velocity of propagation in vacuo is independent of the wavelength. This does not apply to electron rays, for, besides lowering the wavelength, an increase in voltage also results in a greater velocity of the electrons. This velocity may become as high as io^° cm/sec, i.e., 1/3 of the velocity of light. Since electron microscopy operates with very high tensions, the electrons are "rapid", i.e., rich in energy. At a tension of 57 kV the wavelength amounts to about 5.10"'° cm = 0.05 A (BoRRiES and Ruska, 1939a). This is one twentieth of the wavelength of hard X-rays (about i A) and one two hundredth of the wavelength of soft X-rays (about 10 A). In spite of this extremely small ^^avelength and in contrast to X-rays, electrons have no penetrating power, as the electrons are already totally ab- sorbed by layers of solid substances of a thickness of o.i /x. When passing through an object, they lose part of their energy and leave it with a somewhat smaller velocity, i.e., with a changed wavelength depending on the energy loss in the object. This means that the elec- tron beam, originally monochromatic, becomes polychromatic, and images from electron lenses show not only spherical but also chromatic defects as light microscopic images do. The electron microscope. Since the resolving power of the microscope depends on the order of magnitude of the wavelength of the light used, one might expect great improvement in the resolving power of an X-ray microscope as compared to the ordinary microscope. That dream could not be realized, because lenses for X-rays do not exist. The possibility of focusing electron rays has, however, made the construction of a short-wave microscope feasible. (Martin 1938; BoRRiES and Ruska 1939b; Ardenne, 1940a, b; Zworykin, 1940, 1941 ; Zworykin, Hillier, and Vance, 1941 ; Borries, 1941 ; Induni, I945-) The electron microscope operates according to the same principle as the ordinary microscope. The light source is replaced by a source of electrons. Usually this is a hot cathode, but Induni (1945) has also STUDIES IN GELS 117 constructed an electron microscope with a cold cathode. The electron ravs emitted are focused bv a condenser coil and directed towards the object (Fig. 79). An object coil behind the object projects a real mag- nified image of the object, in the same way as an object coid Cothode~^{~~Z^^^^ Light Source lens in the ordinary micro- scope. In analogy to projec- tion microscopy, this real image is magnified again Obiect Airlock^ and projected onto a screen ^~ \ii Illuminating Optic \r:\l^^^^^^J)=p\ Objective High Tension Supply High Voltage f Unit 40-60 kV L» TT Projective Lens I Final Image ^ {Drying Agents A by a projection coil, com- parable to the ocular. Since electron rays are not visible, a fluorescent screen is used, which lights up in propor- tion to the intensity of the incident irradiation, thus giving rise to a visible image Since photographic plates are not only sensitive to ultraviolet and X-rays, but also to electron rays, the fluorescent screen may be replaced by a cassette for plates if microphotographs are to be taken. The whole path of the rays must lie in vacuo, which is maintained by means of vacuum pumps. For this reason the objects must be in- troduced from the atmosphere into the evacuated apparatus through an air lock. The electron image on the fluorescent screen is observed through a window at the side. As in a projection-drawing microscope, the source of the rays is in the upper part of the apparatus, the object being irradiated from above, giving a projected image at about the height of the table. Fig. 79 gives a comparison with the ordinary microscope according to Induni's description (1945). So long as the objects investigated have a thickness of more than 0.1 fi, the image in the electron microscope is formed in the manner of shadow images. Objects of considerably less thickness (order of Recording Device Fig. 79. Comparison between light microscope (at right) and electron microscope (at left) (from Induni, 1945). Il8 FUNDAMENTALS OF SUBMICROSCOPIC MORPHOLOGY I magnitude o.oi /< = loo A) transmit electron rays. In this case the imao-e formation is due to the fact that many of the electrons are de- fleeted from their rectilinear trajectory by the atoms in the object, in much the same way as a small celestial body which enters the sphere of attraction of a star. Now if the object lens possesses a small aperture, the electrons which are deflected do not reach the image and the object appears darker than the background. Since heavy atoms deflect elec- trons more strongly than do light ones, metallic colloid particles appear darker than organic particles, which often furnish a very faint contrast. It is possible to enhance the contrasts by introducing heavy atoms such as iodine (Husemann and Ruska, 1940), osmium (OsOJ or tungsten (phosphotungstic acid; Hall, Jakus and Schmitt, 1945) as "electron dyes". It must be emphasized that the comparison with "dyes" is not strictly correct because the absorption of electrons is very slight. If there is appreciable absorption, e.g. in thick sections, organic objects are instantly burnt by the high energy released by the captured electrons. Therefore, preparations for the electron microscope must be so thin that the electron absorption is negligible. As indicated above, the contrast observed is due to scattering. The electron scattering comprises different phenomena. In the lirst place there is the coherent diff'raction of the beam in much the same way as in the ordinary microscope. The coherent light of the diffracted rays is apt to interfere and to furnish a uniform image when these rays are collected by a lens. However, the scattering of electrons which causes the contrast in the electron microscope is incoherent, i.e. the deflected rays are no longer able to interfere with each other and to be focused at the proper place in the image screen. Most of these aberrant electrons are scattered elastically, when the ray is deviated by some atom nucleus without loss of energy. But there is also inelastic scattering whereby the electron loses some of its energy, and then not only is it deflected from its original path, but its velocity is slowed down at the same time, so that the wavelength of the ray is increased. This corresponds to a chromatic error. The geometric and chromatic aberrations of the scattered electron cause an indistinct blurred image if they reach the objective. They are therefore screened off" by a narrow diaphragm (Fig. 80); consequently, the more incoherently electrons are scattered by an object, the darker it must appear on the image screen owing to this loss of electron light. In order to obtain 4 STUDIES IN GELS 119 highly contrasting images, the aperture of the objective lens must be as small as possible. On the other hand, a small aperture is unfavourable to the resolving power of the microscope, for, according to Abbe's theory of image formation, the resolving power increases with the aperture and reaches a maxi- mum when this becomes ^-^ i. It would be useless, however, to make high aperture electron lens- es because their lack of correc- tion would produce imperfect images, in the same way as un- •corrected lig-ht lenses. Thev •could only be improved by cut- ting down the aperture, but this would reduce the resolution. The present quality of electron lenses can be compared to that -of the optical lenses at the time Fig. 80. Electron scattering by a specimen and selective effect of the objective Jens aperture (from HiLLiER, 1946). when Abbe began to eliminate their spherical and chromatic defects. The necessary screening of the scattered light and the defects of the lenses require very narrow bundles of electrons with apertures of only 0.00 1 to 0.005. As a result of the small apertures a applied, the Tesolving power d is not as large as could have been expected from the exceedingly small wavelength. As calculated for the ordinary microscope (probably Abbe's theory cannot be applied without altera- tions to the electron microscope, but, curiously enough, the results are plausible), the resolving power \s l\a = 0.05 A/0.002 = 25 A. This minimum, however, is only seldom reached. Usually the resolving power amounts to about 50 A (Kinsinger and co-workers, 1946). This is near to the smallest gold particles which have been demonstrated in the ultramicroscope (60 A). Instead of luminous points, however, true images are obtained. Thus the electron microscopy covers the whole field of particle sizes in colloid chemistry, completing this science by the new branch of colloid morphology. Considering the hard work needed to increase the resolving power of the ordinary microscope from dry systems with d = o.^ fxto quartz immersion for ultraviolet light with d = o.\ /u, we cannot sufficiently express our I20 FUNDAMENTALS OF S U BMIC RO S C O PI C MORPHOLOGY I admiration on realizing how the resolving power has been increased by a factor of about a hundred by the discovery of the electron microscope ! The small aperture of the objective coils is responsible for the great focal depth of electron optical images. This depth determines the ratio between the layer thickness in the object imaged sharply and the re- solving power. In the ordinary microscope with large aperture the focal depth is only about i, which means that a section of several [x thickness can be analyzed into successive optical sections at different levels by means of the fine adjustment. In the electron microscope this ratio is about looo. This is a drawback in the spatial analysis of the object, but it is a very valuable aid to the sharp focusing of the image and to obtaining stereophotographs (Ardenne, 1940b; Muller, 1942a; Heidenreich and Matheson, 1944). The similarities and the dissimilarities between ordinary and elec- tron microscopy are listed in Table XIIL TABLE XIII PROPERTIES OF ORDINARY AND ELECTRON OPTICAL IMAGES Light rays Electron rays Wavelength 8000-2000 A About 0.05 A Penetrating power Fairly great Small Contrasts are caused by Absorption Electron scattering Contrasts enhanced by< Staining Dark field illumination Impregnation by heavy atoms (J, Wo, Os) Shadowing Focal depth Small, ca. i Large, ca. 1000 Most favourable magni- fication Up to 1500 Up to 30,000 Resolving power 3000 A 30 A The electron microscope may be changed into an apparatus produ- cing electron diffraction spectra if the objective current is turned oft", and the projective lens is removed (E. Ruska, 1940). The electron diffraction diagrams obtained have the appearance of X-ray diagrams ; they are only formed if a crystal lattice is present in the object. STUDIES IN GELS 121 Technique of making preparations (Ruska, 1939; Wyckoff, 1949). The penetrating power of electron rays being small, it is difficult to find adequate specimen holders. The most suitable holders are nitrocellulose films of submicroscopic thickness. These can be made by spreading a drop of a collodion solution in amyl acetate on water, which is saturated with this solvent. After evaporation of the amyl acetate, a nitrocellulose film, which in favourable cases is only lo m/z, is left on the water surface. When investigating suspended objects (bacteria, viruses, colloid particles), a drop of the suspension is left to dry on the specimen holder. Only dried objects can be placed in the apparatus, because the exposure has to be made in vacuo. This rules out the observation of living organisms in the electron microscope, and it is also impossible to image cytological objects in their natural swollen state. The methods described are appropriate for the investigation of corpuscular colloids. But in general those methods are unsuitable for reticular colloids with a coherent structure and a different microtechnique had to be devised for these objects. Sometimes gel solutions or gels can, admittedly, be dried on a specimen holder to be imaged (Fig. 86a, p. 128). Before it was possible to prepare sections thin enough for the electron microscope, like those obtained in ordinary microscopy by means of microtomes, all kinds of expedients had to be resorted to. In some cases the thin edge of a wedge- shaped section is thin enough for use in electron microscopy, but no images of suitable dimensions can be obtained in this manner. Thick objects, such as cell walls and fibres, can be teased into small fragments after being allowed to swell (Wergin, 1942), or else they can be crushed into sub- microscopic splinters in a vibrating ball mill. These splinters are suspended and finally dried on the specimen holder (Hess and co-workers, 1941). Some gels can be divided into submicroscopic flocculates by means of special vibrators (O'Brien, 1945). Ultrasonic waves have proved to be particularly suitable for this purpose; e.g., by this method microscopic fibres can be disintegrated into submicroscopic fibrils (Wuhrmann, Heuberger, and MiJHLETHALER, 1 946) without damage to the structure such as the vibrating mill inflicts. Several dif^culties arise when these preparations are irradiated. The electrons absorbed impart a negative charge to the object, resulting in repulsive forces between the structural elements, and may cause inflation of the fibrils at the points irradiated. The changes brought about by this effect, however, are as a rule less striking than those suffered by the object as a result of the heat evolved. On absorption, the great energy content of the fast electrons is mostly converted to heat. Silver and gold can be fused together or even melted completely in the electron microscope. Obviously, therefore, organic compounds become charred if exposed too long. Many objects, such as bacteria, appear brownish after exposure in the electron microscope, even if precautions are taken to protect them. Naturally, the thicker the object, the greater is the heat evolved. The object is not easy 122 FUNDAMENTALS OF S UB M I C RO S C OP I C MORPHOLOGY Film to cool, because in vacuo heat cannot be transferred by convection. All that can be done is to withdraw as much of the heat evolved as possible by means of the metallic ring lying on the specimen holder. The best way to do this is to place a fine wire netting over the ring and to irradiate the object through the meshes. Only when an organic preparation is thinner than o.i /x, does it become sufficiently transparent to electron rays and can be irradiated for some time without damage. Therefore, the aim is to produce sections lo-ioo times thinner than those used in histology. Various microtomes have been employed for this purpose. Claude and Fullam (1946) produced sections of 0.3-0.6 IX thickness with a special rotating high speed microtome. Bretschneider (1949a) arrived at o.i /z with the rocking microtome. Similar results have been obtained by Danon and Kellenberger (1950). The fine 0.1 ju. movement of the specimen holder of the microtome is handi- capped by the imperfection of the micrometer screw, but it can be achieved by the thermal expansion of a massive metal block which has previously been cooled down by dry ice (Newman and co-workers, 1949). Special devices for the block advance on an inclined plane seem to be coming into general use (Hillier and Gettner, 1950). Thin sections of organic materials do not show much contrast in the electron microscope, as their constituents C, N and O produce the same electron scattering as the carrier film. Only cell components which contain phosphorus or which are minerahzed appear to be darker. In certain cases the contrast can be enhanced by osmium fixation of the cells and by staining with phosphotungstic or phosphomolybdic acid. The best contrast is obtained by the method oi metal shadowing, developed by Wyckoff (1949). In a vacuum bell jar a small amount of metal is vaporized and deposited obliquely on the preparation (Fig. 81). As a result the faces of the specimen turned to the source of metal vapour are coated with metal, whereas the opposite faces are not. Behind the object there is a zone free of metal which is called the shadow of the spe- cimen. From this shadow the height of the object can be cal- culated if the shadowing angle is known (Muller 1942b). When a preparation like this is irradiated in the electron microscope, the electrons are greatly scattered at the places where metal has accumulated, passing freely through the zones of shadow. As a result, the picture on the projection screen exhibits an astonishing three-dimensional effect, creating the impres- 0.00211 0.01 u O.OOitn Specimen Fig. 81. Shadowing of a specimen by deposition of metal : s length of the shadow, b height of the spe- cimen, a shadowing angle, h = s tan a. 4 STUDIES IK GELS I23 sion that the objects are obHquely illuminated. On the photographic negative the shadows are black, comparable to the shadows in a land- scape cast by the sun, which is the reason why this method has been called shadowing. Because this effect is very striking, the negatives of shadowed preparations are reproduced and not the positives, as in ordinary photo- graphy. This means that a positive film must be made of every photograph before prints can be made. Metal shadowing permits even very flat objects to be pictured, for the shadow can be accentuated by lowering the shadowing angle. A suitable angle is ^\^ (1:6), furnishing pictures reminiscent of sunset or sunrise illumination with its very long shadows. As we are accustomed to illuminate relief maps from the left-hand top corner, shadowed electron micrographs ought to be oriented so that their shadow points towards the bottom right- hand corner. Only thus do we get the natural impression of a high-relief. If such a picture is turned upside-down, the impression received is of reversed rehef, all elevations seeming to be depressions. Wyckoff (1949) has found the most suitable metals for shadowing to be chromium and palladium. The higher the atomic number of the element, the thinner is the metal film yielding the same effect when deposited on the preparation. Whereas the thickness of a chromium film must be 40 A, a palladium film of 20 A will do. In this respect uranium would be better still. As the scale of pubHshed electron micrographs varies from i : looo up to 1:100000, it is as well to mark the magnification on every individual picture. This is done by putting a black line on the micrograph which represents the length of i /m; for a magnification of 10,000, its length is i cm. Results of electron microscopy. The improvement in the resolving power for structures invisible in the ordinary microscope is most evident from the electron optical images which have been obtained from the silica wall of the diatom Pleurosigma angtilatum (Fig. 82a), the well-known test object for the immersion objective of the ordinary microscope. In the latter case the best objectives show three inter- secting systems' of lines (Fig. 82b), which at the utmost give a vague impression of a perforation (Fig. 82c; Ardenne, 1940b), whereas in the electron microscope Fig. 83a is obtained. The surmised pores are clearly imaged with sharp edges ; and, being so far apart, it is evident from this "coarse" structure that the electron microscope is able to resolve exceedingly minute details. It is shown that the pores do not represent cyUndrical canals running through the silica walls, but that the outer opening is in the form of a slit, while the inner one is elliptic and closed by a sieve membrane. Stereoscopic pictures moreover betray, not canals, but spacious caverns, whose outer openings re- 124 FUNDAMENTALS OF SUBM I C R O S C O P I C MORPHOLOGY . .^m^^M^umm^ir^s c) .a) Pleurosigma angtdalum^' . Sm. contour (from Husted, 1930);/?) light microscopic image scale 1500:1 (from Michel, 1940); c) light microscopic image with numerical aperture 1.4, image scale 10,000:1 (from Ardenne, 1940b). Fig. 83. a) Electron microscopic image of Pleurosigma, image scale 100,000:1; l)) sketch of the spatial organisation of the silica wall; image scale ca. 60,000:1 (from MtJLLER and Pasew.\ldt, 1942). ^) f b) 4 STUDIES IN GELS I25 present the slits shown in the image (Fig. 85 b, Muller and Pase- WALDT, 1942). Hence, the diatom wall is not a massive structure, but consists of an outer and an inner lamella, separated by sub- microscopic spaces and connected by pillar-shaped buttresses (Fig. 85b). Fig. 82c shows what was meant in Table XIII (p. 120) by "most favourable magnification". A microscopic image or a microphoto- graph can be magnified at will by projection, so that the magnification, or better the image scale, does not provide an unambiguous reference by which to compare different microscopes. Nevertheless there is a limit to the magnification of images, in that the contours become rague when the image scale becomes too large. For this reason there exists a "profitable" magnification which is best maintained in micro- photography and which is designated as "most favourable magnifica- tion". Strong magnifications of the microphotographic negatives ob- tained result in poor definition as shown in Fig. 82c, where the systems of lines are hazy as a result of a magnification of 10,000, which is seven times the "profitable" one of 1500. The most successful objects of research for the electron microscope are the submicroscopic particles of suspensoids, such as inorganic coUoids, virus particles, bacteriophages, organic macromolecules which exceed 50 A diameter. Unicellular objects such as diatoms and bacteria are too thick; they furnish black shadow pictures and details are only to be seen if the object is perforated or provided with surface appendages (cilia, flagella). The colloid particles, however, are thin enough to transmit electrons, producing real so-called phase images. Fig. 84a shows shadowed macromolecules of haemocyanin from the blood of a snail. This micrograph was the first clear-cut picture of protein macromolecules (Williams and Wyckoff, 1945). Ac- cording to SvEDBERG, the globulat proteins aggregate by 2, 4, 8 etc. to form bigger particles. This rule (see p. 141) found by experiments with the ultracentrifuge, is now substantiated by electron micrographs such as Fig. 84b (Polson and Wyckoff, 1947). The agents of virus diseases have been found to be macromolecules of different shapes. The classical tobacco mosaic virus is rod-shaped, as proved by indirect methods (double refraction of flow. X-rays). The electron micrograph (Fig. 84c) shows that the length of the rods is not defined. Their mean length depends on the p^^ of the dispersing 126 FUNDAMENTALS OF S U B M I C RO S C OP I C MORPHOLOGY Fig. 84. Electron micrographs of globular colloids, a) Macromolccules of haemocyanin (Williams and Wyckoff, 1945). b) Macromolecules of haemocyanin aggregated in fours (PoLSON and Wyckoff, 1947). c) Tobacco mosaic virus (Wyckoff, 1949). d) Cn'stallizcd tomato bushy stunt virus (Wyckoff, 1949). STUDIES IN GELS 127 medium (Takahashi and Rawlins, 1948). There is some indication that the rods are formed by linear aggregation of" roundish particles. The diameter of the straight rods is 150 A (Wyckoff, 1949), which corresponds to the lateral identity period of 1 5 2 A revealed by X-ray investigation (Bernal and Fankuchen, 1937) in the hexagonal crystals of the virus protein (Stanley, 1935, 1936). Contrary to expectation, virus diseases with rod-shaped particles are rare, globular virus macromolecules occurring much more fre- quently. When dried, a virus suspension of this kind crystallizes, and Wyckoff (1949) succeeded in producing very beautiful pictures of the lattice of those crystals (Fig. 84d). The arrangement of the molecules revealed by the X-ray diffraction method can now be seen, and it is most interesting to observe how frequently small disturbances within the regular pattern of the molecule arrangement occur. Figs. 84a-c represent the dispersed particles of protein sols which prove the applicabihtv of the electron microscope in biochemistry. The biologist asks, therefore, what information the electron microscope may give on the structure of gels, among which we classify the shaped portion of the protoplasm. By way of example we reproduce in Fig. 8 5 the electron optical image of a V2O5 gel serving as ultrafilter (Ardenne, 1940b). One recognizes the reticular structure assumed on the basis of results obtained by indirect methods. The agreement with the scheme of Fig. 53a (p. 69), proposed before the electron microscope had been dis- covered, is most striking. Fig. 85 is a rather indistinct picture of a dry gel. The first clear-cut elec- tron micrograph of a very loose gel which, previously to preparation, contained about 99*^0 water, is reproduced in Fig. 86a (Frey-Wyss- ling and Muhlethaler, 1944). It displays a beautiful spatial frame- work with big meshes and roundish interstices. Fig. 86a seems Fig. 85. Electron micrograph of an ultrafilter of vanadium pentoxide, image scale 35,000: i (from Ardenne, 1940b). 128 FUNDAMENTALS OF S I' BM I C R OSCOP I C MORPHOLOGY ^*} If ivvS''- Vi'.'fVV'lAfK ^mM ^ ..^ wt^^ •?/-^ Fig. 86. Electron micrographs of gels, a) Gel of V2O3, 40,000 : 1 (from FREY-\X'YSSLiNGand MuHLETHALER, 1 944). Z)) Cellulosc Cell wall of the alga ijo/rogyra, 19,000:1 (phot. A. Vogel). f) Chitinous cell wall of the fungus Phyconiyces, 12,000:1 (from Frey-Wyssling and MiJHLETHALER, 1950). (7)Tunicin of the mantle of Cioiia, 16,000: i (from Frey-Wyssling and Frey, 1950). 4 STUDIES IN GELS I 29 to contradict the principle of short-range order, .,s the gel strands show a criss-cross random arrangement, but we have to remember that the picture represents a projection of the reticular texture, because the great focal depth of the electron microscope causes gel strands separated in space to be imaged in a single plane. It is likely that the filaments crossing each other are not lying at the same depth in the gel, but that the majority are oblique with respect to the image plane, as is apparent from the faintness of outline of numerous strand "ends". A stereoscopic view of Fig. 86a justifies the comparison of a gel with a wad of cotton wool. At various points ramifications of the gel strings are visible, showing that, notwithstanding the apparent criss-cross arrangement of the gel strands, there exists short-range order. The figure further shows that, in the case of a gel thickness corresponding to an ultrafilter, all possible orientations occur in spite of the short- range order, so that there exists statistical isotropy, as indicated in Fig. 5 3 (p. 69). The curved micellar strands which are visible in Fig. 86b are particularly interesting because they favour branching of the strands (cf. Muhlethaler, 1949). As shown by Fig. 58 (blood fibrin, Wolpers and Ruska, 1939), the reticular structure postulated has also been found in biological gels. It can also be observed in gels of bacterial cellulose (Frey-Wyssling and Muhlethaler, 1946), where we found cellulose strands of about 250 A diameter. Later the same strands were discovered in cell walls (Frey-Wyssling, Muhlethaler and Wyckoff, 1948). Fig. 86b shows the growing tip of the cellulose wall in the end cell of a thread of the alga Spirog^ra. Whereas the gel strands of vanadium pentoxide (Fig. 86a), due to the atomic number 23 of ^^V, produce sufficient contrast in the electron microscope, the cellulose strands with ^^C must be shadowed to produce distinct micrographs. Figs. 86b-d show how well high- relief pictures of gels can be obtained if they are properly prepared and shadowed. An important prerequisite to obtaining such results is the complete removal of any incrusting material. In contrast to VgOg, biological gels are not only full of water, but also incrusted with all kinds of amorphous substances, such as hemicelluloses and lignins in plant cell walls or proteins in animal skeleton materials. Thus Fig. 86d represents tunicin (Tunicate cellulose) from the mantle of Ciona. All accompanying substances have been removed, so only the strands 130 ir.NDAMEKTALS OF SUBMI C RO SCOPI C MORPHOLOGY 1 of tunicin are left. The chitin of fungi cell walls has been prepared in a similar way. Only after repeated boiling of the objects in io% KOH is the texture of the gel disclosed as in Fig. 86c, where two different textures (parallel texture and dispersed texture, see p. 95) are portrayed side by side. If such methods are not used, many of the biological gels furnish the picture of a homogeneous film, because the incrusting substances have the same electron-optical behaviour as those of the gel framework. Clear-cut micrographs of gels can only be obtained if all incrusting substances are carefully removed. This is a handicap in the electron microscopy of protoplasm, as its frame substances are far less resistent to chemical agents used in purifying the framework of gels than are cellulose and chitin. e. Summary Gels with reticular structure are characterized by the existence of a framework whose constituent parts occupy definite mutual positions. The frame strands have either submicroscopic or amicroscopic dia- meters. In the first case they can be detected by electron microscopy and the submicroscopic morphology of such gels is thus accessible ta detailed direct investigation (Fig. 86). In the second case the framework is formed by chain molecules which cannot be solvated completely and maintain c^txAn junctions. If these junctions are released, the network character is lost. In this case the reticular gel, which originally showed only limited swelling, can change into the sol state via the gel solution. As will be obvious from this definition, there exists a transitional state between the reticular and the corpuscular dispersed state. It will require further studies to elucidate the morphological properties of such gel frames and the nature of the bonds in the junctions, which may be quite different in character (see p. 145). II. THE FINE-STRUCTURE OF PROTOPLASM The great conquests in the field of structural chemistry have been real- ized by means of analysis and synthesis. Analysis provides information about the structural units and, with the aid of synthesis, their position in the molecule is determined. Although no inner relationship seems to exist between chemistry and morphology, i.e., between our know- ledge of matter and that of shape, this same procedure has been the method of research in morphology : detecting the structural units by analysis and determining their mutual position. The latter can be done by direct means both in the macroscopic and the microscopic domain and thus has no need of the indirect methods used in organic chem- istry. However, for the elucidation of the invisible submicroscopic struc- ture of protoplasm, in so far as it is not yet accessible to electron microscopy, analysis must again be combined with some kind of synthesis. It is true that this is not a matter of synthesis in the sense of organic chemistry. We can do no more than unite the structural units obtained by analysis in a scheme which enables us to explain the optical and physico-chemical properties of protoplasm. Because of the exceedingly complicated state of the inner morphologic structure of living matter, only a very incomplete solution of the problem is possible in this way. In this situation one might be tempted to abandon the wearisome road of analysis and synthesis and simply accept protoplasm as a given substance. This is, however, impossible for morphology as a branch of the exact sciences. For, so long as there are possibilities of research, morphology must from an inner necessity continue the analysis of living matter — even the sacredness of the human body failed as a taboo in former times. It is only when all the possibilities of analytic dissection which the human mind places at its disposal have been exhausted, that morphology will bow in awe to the secrets of nature. 132 FINE-STRUCTURE OF PROTOPLASM § I. Cytoplasm II a. Molecular Constituents of the Cytoplasm The chemical composition of the cytoplasm is described here only from the point of view of the molecular shape of its compounds (Sponsler and Bath, 1942). The molecular structures concerned are known in principle, but an attempt at morphological synthesis of cytoplasm with the aid of these structural units is impossible. Never- theless, this morphological point of view enables us to explain the physico-chemical behaviour of cytoplasm to a certain extent. Proteins. The basic substances of the proteins, isolated by means of hydrolysis and paper chromatography, are a-amino acids which possess the structure given in Fig. 87a, where R represents a group of C-atoms. a) R CHNHz-COOH cc rRChf- To be exact, the NH2- and COOH- ^NH ,NH2 basic RCH •^ b) '^CQOH O'^'cf /vrt?;. \ CHR NH CO -RCH XO ^o i XHR- ^NH ".CO i HOOC Fig. 87. Molecular structure of amino acids, a) Overall formula; h) principle of chain formation; c) polypeptide chain. groups should be bound to the C-atom as individual atom groups, as shown in Fig. 87b. It can easily be seen that two amino acids can form a so-called dipeptide by eliminating water. If this process is repeated many times, a long polypeptide chain is formed, the ends of which have been left open in Fig. 87c. Like the paraffin chain, it is kinked. The distance between two equivalent groups is 3 . 5 A, as has been ascertained by means of X-ray analysis of crystalline fibre proteins. Only the >CO and >NH groups are similar along the whole length ofthe chain, while R differs ac- cording to the kind of protein and thus is responsible for the great variety in this class of substances. The zig-zag chain drawn in Fig. 87c can be considered as a relatively indifferent frame, which cannot be responsible for the chemical lability which we know the cytoplasm to possess. Its unusual reactivity is due to the side chains R. In chemical text books the amphoteric character of the proteins is often explained by the fact that amino acids possess both an acid and a basic group (Fig. 87b). However, it follows from the structural picture of the polypeptide chain that these groups disappear in the condensa- tion process, thus losing their capacity for dissociation. If in spite of CYTOPLASM 135 this the proteins clearly show acid or basic properties, this is brought about by the side chains which in their turn carry free CO OH- or NHg-groups. This happens when some members of the polypeptide chains consist of dicarbo-amino acids or diamino acids (Fig. 88). Acid side chains NH ^CH-CH^-COOH CO / NH \ /' CO Aspartic ocid CH-CH^-CHp-COOH Glutamic acid Lipophilic side chains ,CH-CH,-CH / ^CH- CO L Pucine / NH ycH-cH,^^ CO ^ Phenyl alanine Sulphur containing side chain / NH yCH-CH.-SH CO Cysteine Basic side chains NH \ ^CH-CH2-CH^-CH2-NH2 CO Ornithine ^LH-CH.-CH.-CHj-NH-C^ c'o ^^^ ^ Arginine (Valine -^ Guanidme) Hydrophilic side chains NH .iH-CH^-OH CO Serine \ / NH \h-chp-(^oh CO ^ Tyrosine Possible chain end CH,^CH2 NH \ \ ^CH! ^•P Proline Fig. 88. Side chains R of the polypeptide chains. The common amino acids (valine, leucine, phenyl-alanine, etc.) cannot exercise special influence on the reactivity of the proteins, but they confer upon these proteins a pronounced lipidic character, since the ends of the side chains consist of methyl or phenyl groups (Fig. 88). In many cases, however, the terminal groups carry an alcoholic hydroxyl group (serine, tyrosine), on account of which a certain hydrophily is maintained. A particularly important side chain is cysteine with its very reactive sulphydryl group. As will be shown later, this group very easily forms bridges between neighbouring polypeptide chains. In contrast to such constituents of protein chains, capable of bonding and thus favouring further polymerization, cyclic amino acids such as proHne can ter- minate the main valency chains and thus limit the apparently endless 134 FINE-STRUCTURE OF PROTOPLASM II polypeptide chain molecules. The proline ring can, however, also be built into the peptide chain (see Fig. 173, p. 546). Considering the variety and the number of 20 amino acids (besides some rare amino acids, Cohn and Edsall, 1943) which have thus far been isolated from proteins, and in view of the fact that these can occur as side chains at various points along the polypeptide chains, we realize that the protein components of the cytoplasm represent a variegated mosaic. It follows from Fig. 88, that the amino acid configuration -CH-NHg-COOH does ;w/ contribute to the character of the mosaic, since it is only responsible for the peptide interlinking. The chemical behaviour of the polypeptides of protoplasm is determined by the end and side groups of the amino acids, to which often little attention is paid. The polypeptide chains show a number of properties which single them out from the other substances of which protoplasm is built up. 1. The principle of repetition which mhlology we know as segmenta- tion or metamerism. Most high polymer substances are built according to this principle. In the majority of these substances, however, iden- tical monomer groups are repeated, whereas in the polypeptide chains the side groups R, which occur at regular distances of 3.5 A, have different constitutions. Probably the typical side chains also repeat themselves regularly, but their period is much greater and is often not accessible to experimental analysis. 2. The principle of specificity. Owing to the numerous possible side chains R and the unlimited variety in their arrangement along the polypeptide chains, an almost infinite number of polypeptides is con- ceivable, distinguished only by slight difference in construction. This difference in construction may result in a different chemical behaviour which becomes apparent in the specific properties of the proteins. 3. The principle of contractility. The most striking property of poly- peptide chains is their capacity to contract, as will be further discussed on page 559. The origin of the mobihty of cells (protoplasmic flow, cilia, contractible fibrils, etc.) must be sought in these molecular structural units and for this reason they form undoubtedly the most important structural elements in the fine-structure of protoplasm. The number N of amino acids in natural polypeptides seems to obey the Bergmann-Niemann rule (1936/37) N = 2" -3™, which in- dicates that there must be some threefold symmetry in protein mole- CYTOPLASM 155 :h, ch. NH-, SH OH CH3 cules. Established N-numbers are 96 = z% 144 = z*-f, 288 = z*-^^, ^^^ ^ 2«-32 etc. (ScHEiBE, 1948). These complicated proportions have been derived from crystalline proteins. In this respect there exist two different types : a) Globular proteins consisting of isodiametric macromolecules which preferably crystallize in the system of cubic, hexagonal or •orthorhombic closed packing (Fig. 90a). b) Fibrillar proteins formed by expanded polypeptide chains aggre- gated to a chain lattice (Fig. 90b). In the chain lattice they may assume a spiral configuration (spiral chains, Perutz, 195 i). In the second type the crystallization depends on the regularity of the side chains R. If these side radicals are simple as in silk fibroin (Fig. 170), where they consist mainly of H- and CHg-groups, the chains combine as easily as polysaccharides to form a crystal lattice. As will be obvious from Fig. 89, however, this is not possible if the side chains happen to be of quite different lengths and confio-urations. These con- ditions can be compared with the arrangement of bean- or peastalks. Where- as there is no difficulty in uniting a great number of smooth bean stalks into a btmdle, it is not so easy to obtain a parallel order in pea stalks with their numerous twigs pointing sidewise; and if, moreover, the lengths of these twigs alter- nate in an irregular manner, the resulting structure becomes so spaci- ous that it is almost impossible to bundle them together. This is the case with complicated polypeptide chains. In general these unwieldy chains are folded up in some complex manner to form globular molecules. Open spaces inside these macro- molecules are occupied by bound hydration water. The protein part- icles crystallize in a molecular lattice of close packing. As their size is considerable, some space accessible to additional water or even dye- stuff molecules is left between the spheres (Fig. 90a). Such crystals therefore swell or shrink and can be stained in aqueous solutions. Rigorous dehydration removes not only the water between the CH. COOH NH2 NH Fig. 89. Unequallengths of polypeptide side chains R. 136 FINE-STRUCTURE OF PROTOPLASM 11 macromolecular Spheres, but also the hydration water inside the glob- ular molecules, so that their structure is destroyed and the solubilitv of the protein is abolished. This physico-chemical transformation of soluble proteins is called denaturation. There are some indications that the denaturation of globular proteins consists in an unfolding of the ? «) 1/ / \ V) Fig. 90. Model of the fine-structure of protein (from Frey- Wyssling, 1944b). a) Lattice of spherical macromolecules (slightly anisotropic or isotropic; highly hydrated); b) chain lattice of thread molecules (strongly anisotropic, barely hy- drated). The transformation a -> b is termed "denaturation". wrapped-up polypeptide chains. In Fig. 90 the denaturation of globular into fibrillar proteins is indicated by an arrow a -> b. The inverse reaction, the transformation of the denatured protein into globular molecules, is usually impossible in vitro, but it must occur readily in vivo. Forms of protein molecules intermediate between the globular and chain configurations are not well known. Such intermediate shapes do not crystallize out, but it is probable that they are involved in protein metabolism. Fig. 95c (p. 144) shows the length of the poly- peptide chain which is folded up in a globular protein molecule of 100 A diameter, Ranzi (195 i) has devised a method for distinguishing globular from fibrillar proteins in dilute solutions. The first show an increase in viscosity with KCNS as compared with a test solution equimolar I CYTOPLASM 137 in KCl, whereas the second show a decrease. With this test Ranzi has shown that the euglobulin of frog embryos is fibrillar between p^ 5.5 and 8.5; but beyond this range it is globular. Since the develop- ment of the frog ectoderm in tissue cultures is only possible within the range of pH 5 and 9, there is an indication that fibrillar proteins are indispensable for any manifestation of morphogenesis. Lipids. The biological concept of lipids comprises all substances which are hydrophobic. This concept is therefore characterized by a negative property (insolubility in water) rather than by a positive one :y OH-CH2 y CH3 -c I ~^ §■ OH-CH "I CH3 ■0 I Q. 1^ OH-CH2 -■ CH.f 3 Fatty acids + Glycerol 0 Lipophilic CH3 \y\y\/sy\y\/\/\/\ C-O-CH, /\/\/vv/VN/s,/\/sy\/v/\/'VN/ CW-, Lipophilic Falty acid Higher alcohol Wax Fig. 91. Molecular structure of lipid chains. (solubility in organic liquids). For this reason it comprises different families of substances such as terpenes, waxes, fats, sterines, etc., of which the last two take part in building up protoplasm. True lipids are characterized by the fact that all their free end groups consist of typically Hpophilic groups. This is especially obvious in the case of fats, which represent esters of the three-valent alcohol glycerol with fatty or oleic acids. As a result of the esterification, thehydrophilic groups of the original products are screened, as shown in Fig. 91. In the same way the hydrophilic groups are masked in waxes which are formed by the esterification of higher alcohols with higher fatty acids (Table XXVII, p. 296). It is difficult to say why they are screened in the course of the metabolic process, but in any case these lipids contrast strongly with the hydrophilic compounds of living cytoplasm and, if they are formed in excess, we observe the well-known phenom- enon of fatty degeneration of protoplasm. A correct balance between hydrophilic and Hpophilic compounds in living matter is essential. In contrast to fats, most lipophilic compounds of the cytoplasm carry at least one hydrophiUc group, which serves to bring about the 138 FINE-STRUCTURE OF PROTOPLASM II Fie. 92. Molecular structure of cholesterol; terminal group OH. contact with neighbouring hydrophilic groups. This applies in partic- ular to the important group of the sterines (Windaus, 1923), from among which the formula of the complicated cholesterol C27H45OH is reproduced (Fig. 92). The molecule contains four rings and a double bond. According to X-ray ana- lysis (Bernal, 1932) the length of the molecule is 17-20 A and its cross-section only 7.2 A. The elongated form tallies well with the optical finding that dissolved cholesterol molecules can be easily oriented in a field of flow and like most rod-shaped mole- cules show positive birefring- ence of flow. Phosphatides. Because of their solubility in ether, phosphat- ides are usually also counted among the lipids, but besides their lipid character they possess a marked tendency toward hydrophily, which is shown by their adsorption of water and the occurrence of myelin forms. Thus, phosphatides represent compounds which are intermediate between hydrophobic and hydrophilic substances and for this reason belong to the most important intermediates between the representatives of these two extreme groups in the cytoplasm. By way of example we may mention lecithin which, like o^ the fats, consists partly of glycerol and fatty acids. ^- In this case, however, only two OH-groups are occu- "^ pied by fatty acids, the third being esterified by phosphoric acid and the latter in its turn by the amino alcohol choline (Fig. 93). Choline HOCH2-CH2-N(CH3)30H is a base whose hydroxy 1 group is attached to a methylated ammonium group. The three methyl groups might conceivably give the end group -N(CH3)30H of the molecule a lipophilic character O HO-k=0 0 CH2 CH2 I I 0 0 1 I CHy CH2 H^C CH2 H2C CH2 H2C : j \h2 H2( CH2 H2C CH2 H2C H.C CH2 O <\ o CH2 CH2 H2C HpC CH2 CH2 H2C H,C CH, CH2 H2C H.C CH, \ Lecilhin Fig- 93- Molecular struc- ture of lecithin. I CYTOPLASM 139 in spite of the hydrophilic OH-group. This, however, is not the case. For, curiously enough, alkyl groups (-CH3, -C2H5) bound to am- monium nitrogen are hydrophilic in behaviour (like methyl bound to oxonium oxygen, which makes pectic acid and methyl cellulose soluble in water, seep. 60). For this reason the ammonium end group tends to escape from the neighbourhood of the lipophilic end groups of the fatty acids. Consequently, the lecithin molecule resembles a tuning fork (Fig. 94), in contrast to fats which can be represented schematically by a three-pronged fork without a handle. The prongs of the fork represent the lipophilic pole, the handle of the fork the opposite hydrophilic pole of the lecithin molecule. The phosphatides react with the protein chains of the cytoplasm by combining with either the lipophiHc or the hydrophilic end groups of the side chains, as indicated in Fig. 94. This junction is not of a o ° o ^oOH- o o 000 ■OH OH- ■ CH- K 000 '-"j > 00 CHy -CH3 CH, CH, CH:, CH3 CHs Lecithin Fat Water Lecithin Polypeptide chain Fig. 94. Relation between polypeptide side chains and lecithin; o = water molecule. chemical nature, for the phosphatides can be extracted from the cytoplasm with ether. Nevertheless the phosphatide molecules occupy quite definite places, according to the character of the side groups in the polypeptide molecules. Lipids without hydrophylic groups, such as fats, can combine only with the lipophiHc side groups. For this reason their possible combinations with protein chains are limited. As shown in Fig. 94, they can only enter into relation with hydrophilic side chains by interposition of phosphatides or other intermediates. The sterines possess a polar structure similar to that of the phos- phatides, but lecithin is more reactive: of its two hydroxyl groups at the hydrophiUc tail one is acid (attached to phosphorus) and the other basic (attached to nitrogen). For this reason it can form salts with basic as well as with acid groups of the polypeptide chain. Phos- phatides can therefore react with nearly all end groups occurring in the side chains of proteins. Sterines, on the contrary, are only capable of forming esters. Finally, for fats, all side chains of the polypeptides, with the exception of the lipophilic end groups, are blocked. This I40 FINE-STRUCTURE OF PROTOPLASM II shows clearly how opportunities of entering into the protoplasm multiply as the lipids become more hydrophilic in character. Chemical composition of the cytoplasm. The proportions of the com- pounds described above vary considerably in the cytoplasm (Table XrV). This is especially so for lipids, carbohydrates and water-soluble compounds. Although the two analyses in Table XIV represent ex- treme cases, it is evident that protein is the main constituent of the cytoplasm. Only small amounts of the other constituents are structural compounds. In the cytoplasm of the sUme mould Reticularia there are considerable quantities of reserve substances such as carbohydrates, soluble nitrogenous compounds and probably most of the lipids. Whereas slime moulds can be dried and analysed without difficulty, the cytoplasm of tissues with solid cell walls can not. In this case the membranes have to be broken in a blendor and the cell contents sus- pended in an appropriate solution, from which the constituents of the cell wall, the cytoplasm, the plastids and the nucleus must be separated by centrifuging and by fractionated salting out, e.g. with ammonium sulphate (Menke, 1938a). The fraction corresponding to the cytoplasm yields the analysis recorded in the right column of Table XIV. By this method of preparation, all water-soluble com- TABLE XIV CHEMICAL COMPOSITION OF CYTOPLASM Plasmodium ReticuLria Ijcoperdon KlESEL 1930, p. 257 Leaves Spinacia okracea Menke 1938a, p. 289 Proteins Soluble N-compounds. Lipids Phosphatides Cholesterol Nucleic acids Carbohydrates . . . . Ash Unknown 29.07 12.00 19.05 4.67 0.58 5.68 25.08 5.87 100.00 0/ o 85.0 0.7 3-1 II. 2 1 00.0 CYTOPLASM 141 pounds, such as sugars, amino acids and amides, are lost; hence the proportion of insoluble proteins is increased. It is noteworthy that the proportion of lipids is very low. The considerable "unknown" fraction probably comprises the nucleic acids, which in the present case have not been separately determined. b. Physicochemical Behaviour of Proteins Si^e of globular protein f/wlecules. The molecular weight of globular proteins can be determined with the aid of the ultracentrifuge. SvEDBERG has found that the weights of quite different proteins are similar. Thus the molecular weight of pepsin, insuHn and egg albumin is 34,500 (SvEDBERG, i93i).In many instances there is an approximate multiple of this figure, such as 70,200 for horse serum-albumin. After a systematic investigation, Svedberg (1938b) came to the conclusion that in protein molecules there is a fundamental unit of molecular weight 17,600. TABLE X\' WEIGHTS AND SIZE OF GLOBULAR PROTEIN MOLECULES Substance 0 ci. Q > J3 EC 3 c ^ O O £ 'E I S ° k -^ v-l ^ 1 nt u 1 -C J2 ■l-J °<: 3 0 u u £ particles din A 3 u < n cr -a Q C/3 1 1 Lactalbumin a, rnyoglobin i 17600 200 Lactoglobulin, ovalbumin, zein, pepsin, insulin 2 35 200 400 Serum albumin, CO-haemoglobin, yel- low ferment 4 70400 800 Serum globulin 8 140800 1600 Edestin, excelsin, phycocyanin, phyco- I erythrin, catalase 16 281600 3200 Haemocyanin (cleavage component), | urease 24 422400 4800 Haemocyanin (cleavage component). . 48 845 000 9 600 Haemocyanin (Calocaris) j 96 1690000 19200 Haemocyanin (Rossia) I 192 13380000 38400 Haemocyanin (Helix pomatia) .... 1-384 6760000 j 76800 c c 34.5 2.87 1.8 I 0.07 43.5 5-75, 2.o; 0.14 55 II. 5 2.2 0.28 69 23 2-5 0.56 87 46 2.8 1. 12 100 69 3.0 1.68 125 138 3-4 3.36 158 275 3.8 6.72 200 550 4.2 13.44 250 IIOO 4-7 26.88 j^Z FINE-STRUCTURE OF PROTOPLASM II Table XV gives examples of this multiple series (cf. K. H. Meyer, 1940a, p. 409). It shows how the Svedberg units combine in 2's, 4's, 8's, i6's etc. There are, however, not only multiples of 2, but also of 5 (e.g. 24), a fact which recalls the Bergmann-Niemann rule. Up to 384 units may be combined in one molecule. The aggregation or dissociation of these large particles depends on p^ conditions. Since the nitrogen content of proteins is 16%, the average mole- cular weight of the amino acids in proteins is 6.25 x N = 87.5, if no allowance is made for basic amino acids with more than one N-atom. With this figure, the approximate number of amino acids in globular protein molecules can be calculated. The Svedberg unit contains about 200 (which is near to the figures of 2^ x 3 = 192 or 2^ x 3^ = 216) and the largest particles mentioned in Table XV contain more than 75,000. Globular protein molecules can be photographed in the electron microscope (Fig. 84a, b, p. 126). The average space needed by an amino acid (Fig. 181, p. 365) is 3.5 X 4-6 X 10 A^ = 161 A^. In the electron microscope a sphere of diameter 50 A can be readily recog- nized. Its volume is 50=^ x n/G A^ = 65,500 A^. This corresponds to about 400 amino acids. Protein molecules with two Svedberg units must therefore be easily visible in the electron microscope, while the Svedberg unit itself is just at the limit of the resolving power. A similar result is obtained if we remember (Fig. 31b, p. 34) that in an aliphatic chain the carbon atoms are lined up at intervals of 1.25 A, the distance between neighbouring chains being 5 A. Thus 40 X 10 X 10 = 4000 carbon atoms can be placed in a cube of 50^ A=^. This would yield a molecular weight of 48,000, which, again, corre- sponds roughly to 2 Svedberg units. A third determination is possible based on the average density of proteins, which is 1.33. Knowing the absolute weight of a Svedberg unit (17,600 divided by the Loschmitt number 6.06 X lo^^), the volume of the molecule can be calculated. Considered as a sphere, its diameter is 34.5 A. In Table XV the size of the macromolecules in the multiple series of globular proteins has been calculated in this way (Frey-Wyssling, 1949a), the dimensions found being as shown in Fig. 84a, b (p. 126). As a result we may note that globular macro- molecules of protein with at least 400 amino acids or a molecular weight of about 40,000, are within the resolving power of the electron I CYTOPLASM 145 microscope. If the diameter d is calculated from the mean space of 161 A^ needed by an amino acid, somewhat largervalues are obtained. The length of the completely unfolded polypeptide chain of the denatured molecule is found by multiplying the chain period of 3.5 A (trans-) or 2.8 A (cis-configuration,seep. 346) by the number of amino acid residues in the molecule. With the period 3.5 A, the lengths L indicated in Table XV are obtained. Of course, these figures are maximal values which are not realized, since the chains will never expand completely, but will assume a bent or curled shape. Surface films of proteins. Although globular proteins are soluble in water or salt solution, not all parts of the molecule show an affinity for water. The polypeptide chains which are coiled up in an unknown manner within globular molecules carry hydrophilic and lipophilic (hydrophobic) side groups. The former strive for contact with water but the latter "avoid" it, hence the proteins can be spread as molecular films on the surface of water (Gorter and co-workers, 1955). One milligram of protein can cover a surface from one to more than two and a half square metres; assuming a density of 1.53, this means films of 7.5 to 3 A thickness (Adam, 1941, p. 87). These values show that the molecular film cannot consist of spherical macromolecules, but that these protein globules flatten and uncoil to form protein chains. This spreading of the macromolecule allows all hydrophiUc groups to make contact with the water surface and all hydrophobic groups to turn away from it towards the air. If the surface film is larger than one square metre, it is liquid, i.e. the flattened molecules retain their mobility and may change their relative positions on the water surface. However, as soon as the film is compressed to an area of one square metre, it becomes solid, rigid and insoluble; the molecules lose their individuality and, because they stick closely together, they can no longer be hydrated. They assume the state of fibrous proteins and as such become insoluble. This change of solubility is known as "de- naturation", mentioned on p. 136. Merely shaking a protein solution often suffices to form a foam of insoluble denatured protein. If the molecular weight of the protein is known, the area per mole- cule in a surface film of 7,5 A thickness can be calculated, as has been done in Table XV (p. 141). If this area is considered to be circular, the diameter D of the circle can be compared with the diameter d of the globular molecule. For small protein molecules the ratio D/d is 144 FINE-STRUCTURE OF PROTOPLASM II about two, and for larger ones four to five (Fig. 95 a, b). This means that the area of the flattened molecule is four to twenty times bigger than the cross-section or projection of the spherical molecule before spreading. The polypeptide chain may wind about in this area. If the cross-section of such a chain measures 4.6 x 10 A as in the chain lattice, its length L can be computed. Fig. 95. Surface film of a protein (from Frey-Wyssling, 1949a). a) Globular molecule of 100 A diameter; b) spread to a surface layer 7.5 A thick; c) denatured to a polypeptide chain 11,600 A long. The chain length L obtained for the globular particles is shown in Table XV. For instance, a protein molecule of 24 Svedberg units with a molecular diameter of 100 A harbours a chain of 11 600 A = 1.16 ^ length (Fig. 95 b, c). An even greater length is obtained if it is assumed that this molecule consists of 4800 amino acids, each of which contributes 3.5 A to the chain length; this yields L = 1.68 fi. Since globular proteins denature so easily, we may ask what types of force hold together the inner architecture of these macromolecules? They must be rather weak, because they are broken by mere contact of the globular molecules with a water surface. On the other hand, the expanded molecules form a solid film, which has the character of a fibrous protein. It must be supposed that the individual molecules have been fused to a two-dimensional molecular aggregate. Here, instead of intramolecular forces holding together the coiled, folded or laminated internal structure of the globular molecule, inter molecular forces unite neighbouring expanded molecules. The same thing occurs when globular protein molecules are connected to form beaded chains. CYTOPLASM M5 These bonds between macromolecules are very important, because they transform the protein from the state of a corpuscular sol into that of a reticular gel. But in spite of this fact it is probable that the intramolecular and intermolecular forces are alike, because it makes no difference whether distant parts of one polypeptide chain or sections of two different chains react with each other. In both cases attractive forces between side groups are involved. The places where the side chains are mutually connected will again be called "junc- tions" (see p. 67) and the nature of these points of attachment will now be discussed in more detail. The theory of junctions^. The attraction between the side groups of neighbouring protein mole- cules may be of a number of V different types. Some of these I CH3 ■cHj y CH3 CH3- o°o° o O O n .C/" ^OH OH NH3 SO4 0 -c- -0H° OHf- O O O o o o 0 ^ II CO-NH3 Salt formation -COOH* o '^ o ° -0- Ester bridge W -S S- 0 NH II II C-NH-C- Amide bridge Sulphur bridge -0- Ether bridge possibilities are shown in Fig. 96. Both hpophilic and hy- drophilic groups may attract each other. Salt-Uke or ester- like bonds can be formed be- tween neighbouring acidic and basic or alcoholic groups, and even main valency bonds may be operative, forming ether-, acid amide- or sulphur bridges. Not all side chains take part in these reactions, but a certain number with free end groups will combine with lipids, hydrophiHc groups or water, as has already been described (Fig, 94, p. 1 39). Furthermore, they form points of attraction for ions of the inorganic salts which, according to their charge, will gather round acidic or basic groups. It is important that the end groups of many side chains remain free, for if they were all interlinked, the result would be a molecular aggregate of very small reactivity. There exist four kinds of junctions keeping together the molecules formed by polypeptide chains. In Fig. 96 these have been numbered I-IV, and thev can be characterized as follows : ^ In German: Haftpunkt-Theorie Fig. 96. Schematic representation of junction possibilities between neighbouring polypeptide chains; o = water molecule. 146 FINE-STRUCTURE OF PROTOPLASM II I. Homopolar cohesive bonds, i.e., mutual attraction of lipidic groups ; 11. Heteropolar cohesive bonds, i.e., attraction between groups of pronounced dipole character; III. Heteropolar valency bonds, i.e., formation of salts and esters; IV. Homopolar valency bonds or bridge formation. We shall briefly discuss the characteristics of these types of bonds. I. Homopolar cohesive bonds are of the same kind as the forces which keep a paraffin crystal together. Very little is known about the causes of the attraction between lipophilic groups, for the electric charges in these substances are distributed so regularly that the resulting field of force is negligible, in contrast to dipole molecules. It has therefore been suggested that weak dipole moments are induced in the neigh- bouring molecules by periodic oscillations in the field of force, brought about by vibrations within the electronic configurations (Bartholome, 1956). We know more about the energy of these bonds. As follows from Table IV (p. 32), the cohesion between methyl and methylene groups is the weakest among the cohesive forces. This kind of bond is loosened by small amounts of energy and is therefore strongly sensitive to temperature changes. For this reason, paraffins, fats and waxes melt at relatively low temperatures in spite of their high molecular weight. A similar behaviour is shown by the homopolar cohesive bonds between lipidic side groups of neighbouring polypeptide molecules. By a rise in temperature, this kind of junction is easily loosened. Similarly, lipids and phosphatides which are attached to these groups become more mobile. This causes the living matter to liquefy to a certain extent: the rapidly decreasing viscosity of the cytoplasm as a function of the temperature is a well-known phenomenon (Heil- BRUNN, 1930). Fig. 97 shows the rapid decrease in the viscosity of amoeba cytoplasm between 10 and 20 degrees C, which is probably due to the rupture of lipidic bonds in addition to the viscosity decrease of the intermicellar water. At temperatures beyond 20° C. another process sets in, vi2., a shrinkage at those spots where hydrophilic chain ends come together, resulting in some kind of solidification. At the same time, however, the rupture of lipidic junctions continues and at 25° C. clearly surpasses the solidification brought about by de- hydration. By raising the temperature still further, the curve should CYTOPLASM 147 finally rise again, since in that case the cytoplasm would solidify as a result of shrinkage. Death with coagulation occurs at about 42° C. Since in the physiological temperature range a rise in temperature would certainly not be able to rupture either heteropolar cohesive bonds or main valency bonds, it is permissible to attribute the change v Pf) p A 0 > 20 A "A I i \\ I'i , 10 \ ^ ° r \ 5 \ J \ \/ \ \ b-o-^ i 0 5 K J 7; ) 2( ■) 2t ) 3 0 3 5 °C Fig. 97. Viscosity of the cytoplasm of the amoeba (from Heil- BRUNN, 1930). Abscissa: temperature in °C. Ordinate: viscosity (time in seconds, which a crystal enclosed needs to travel halfway through the cell under the influence of gravity). in viscosity of the cytoplasm primarily to the abolition of homopolar cohesive bonds. The weakness of the homopolar cohesive bond is demonstrated by the exceedingly small surface tension of proto- plasmic membranes (i dyne/cm against nutrient. Table XXI, p. 166), in comparison with water (71.6 dynes/cm against air. Table V, p. 43), where the surface is formed by heteropolar HgO molecules. 11. Heteropolar cohesive hands are of a quite different character. The underlying attractive forces are due to dipole moments (p. 19), which are mostly so strong that they are designated as secondary or residual valencies. Of recent years the semi-chemical character of heteropolar cohesive bonds has come to the fore, since they are designated as hydrogen bonds or hydrogen bridges (Pauling, 1940). Wherever dipolar groups with hydrogen atoms situated in the periphery (OH-, NHg-groups) 148 FINE-STRUCTURE OF PROTOPLASM II are present, the possibility exists of their being attracted electro- statically by the local negative charges of the dipole groups of neigh- bouring molecules. To a certain extent the hydrogen atom acts as an intermediary between the two molecules and connects them by forming some kind of bridge. This is represented in Fig. 98 for two poly- peptide molecules running in opposite di- c^o H- N rections. The hydrogen atom is ifted some- RHC CHR what out of its position in the original N-H o^c molecule and it looks as if part of the hy- O'^c N H- drogen valency is transferred to the neigh- CHR RHC bouring molecule. Clearly, this schematic H-N c-^o representation of the "secondary valencies" 9^'^ >^-i^ gives only a very incomplete idea of the ^HC CHR interactions of the two electric fields which f'"'^' o-^c attract the positively charged hydrogen °""^\ ^/v- "«• • atom with different field strengths. CHR RHC^ j£^ £qj. 5^g£(- reasons, the heteropolar "••■^ "■"\ /'^° - groups (OH, COOH, CHO, NH^ etc.) of „ ^^ , , , , neighbouring molecules cannot come near Fig. 98. Hydrogen bonds be- ° ° tween polypeptide chains. enough together, their electric fields attract water molecules. Instead of hydrogen bridges, a hydration layer is formed between them (Fig. 96, p. 145) and it is obvious that with this kind of junction the cohesion depends on the number of water molecules between the two end groups, i.e., on their hydration. For this reason heteropolar cohesive bonds are j-^/zj-zV/Vf fo hydration changes. Swelling depends largely on the presence of inorganic ions, in which case the so-called ion series of Hofmeister holds good (see HoBER, 1922). Their influence on swelling phenomena can be ex- plained morphologically on the basis of the diameter and hydration layers of the ions. Goldschmidt has calculated the diameters of the ions from the distances between the atoms in the crystal lattice, and the size of the hydration layers can be derived from the ion mobilities. For the monovalent cations, for instance, the following radii have been found (Table XVI). Obviously the small ions have thicker hydration layers than the bigger ones. This is due to the fact that the water dipoles are attracted more strongly as the distance between the centre of gravity and the CYTOPLASM [49 surface of the ion decreases. Fig. 99 shows a graphical representation of the water layers. It demonstrates how the ionic radii grow with increasing atomic weight while the water layers decrease. If a gel swollen in water is imbibed with salt solutions, the pene- trating ions will weaken the electric field of the hydrophilic dipole TABLE XVI ION RADII 1 ^^ Na K NH4 Rb Cs In the crystal lattice according to GOLDSCHMIDT 0.78 0.98 1-33 1.45 1.46 1.66A Derived from the conductivity at 00 dilution 3.66 2.81 1.88 1.89 1. 81 1.80A Number of H^O per ion, accord- ing to Pallmann (1937) . . . 10. 0 4-3 0.9 0.8 0.5 0.2 groups of the gel frame; consequently their hydration decreases, which results in shrinkage. In the case of biogels this effect of shrinkage in neutral salts is observed only in rather concentrated salt solutions (from about N/z upwards) which in most cases must be considered to be non-physiological. Shrinkage by means of salt can therefore be used for preserving purposes (brining of meat) or for the salting-out of dissolved proteins. The degree of shrinkage depends on the radius of the hydrated ions as long as other conditions remain constant. For instance, if dried agar powder swells in Normal alkali chloride solutions (Brauner, 1932), the degree of sweUing is less than in water, and it is found that by comparison with the other alkali ions, Li and Na ions result in a higher degree of swelling, in accordance with the series of Fig. 99 : Fig. 99. Hydration. Ions of the alkali series; hydration layer dotted. 150 FINE-STRUCTURE OF PROTOPLASM II Li > Na > K > Rb > Cs Using only potassium salts and varying the anion in the halogen series, one finds : I > Br > CI, i.e., the more strongly hydrated CI causes less swelling than the lesser hydrated I. This inversion of the influence of ion hydration shows that the influence of the ions on swelling phenomena is determined primarily by their charge. Biogels, such as agar in the present case, usually possess a weakly negatively charged gel frame. For this reason the discharging effect of cations of equal valency is inversely pro- portional to their hydration. The effect of the anions is due to the fact that the discharging cations are accompanied by their gegenions. These lay greater claim to the charge of the cations, in inverse ratio to their hydration. For this reason, the discharge of the gel framework by a given cation accompanied by I ions is less than if it were ac- companied by CI ions. In other words, for a given cation, the less hydrated the anion of the salt is, the greater will be the water absorp- tion of the gel. In many cases, however, gels swell not less but more strongly in salt solutions than in water. This occurs if the gel framework possesses ionogenic groups, as is the case with proteins. For example, the gel frame of gelatin, when imbibed with a neutral salt solution, shows a considerable negative charge as a result of the dissociation of COOH-groups. For this reason cations can be retained by ad- sorption ; their hydration is great- er than the dehydration of the gel framework, caused by the adsorp- ^^-r^' tion of the cations. It is therefore Fig. 100. Hydration. Influence of ions on possible for thedegreeof swelling the hydration of polypeptide chains; A = ^gached in salt solutions tO be difference in swelhng. higher than that m water. Fig. 100 indicates how ions of equal valency can cause different degrees of swelling. Consider an anionic side chain and a hydrophilic OH-group of a neighbouring polypeptide chain. Both possess a 1{0H\^ I CYTOPLASM 151 hydration layer. If a Na ion surrounded by its hydration layer ap- proaches this system, it is held electrostatically, and a hydration equilibrium between the various groups is established. If the Na ion is replaced by a much less hydrated ion like Rb, the latter is able to approach the anionic group more closely because of its smaller hydration layer. This results in a stronger discharge than in the case of the Na ion; the hydration decreases and the neighbouring poly- peptide chains approach each other. An explanation along these general lines becomes more difficult if bivalent ions such as Ca take part in these processes. Since bivalent ions carry two elementary charges, they can discharge negative pro- teins more strongly than monovalent ions. For this reason they usually cause shrinkage of protoplasm (decrease in permeability, in- crease in density and viscosity; Cholodny and Sankewitsch, 1935). In the case of the trivalent ions Fe and Al these effects are still more pronounced (tanning). One speaks, therefore, of a valency rule of shrinkaee, which states that the shrinking effect of ions increases with rising valency. With increasing charge of the ions, however, the hydration layer also increases. The Ca ion, for instance, is hydrated more strongly than the K ion of the same size. Accordingly, CaClg causes gelatin to swell to a greater extent, and this can even result in the formation of a sol. In the same way the strongly hydrated Zn ion in concentrated ZnClg solutions causes unexpectedly marked swelling of cellulose. The valency rule does not, therefore, apply generally to bivalent ions. The valency rule asserts itself more clearly in Hofmeister's series of the anions SCN > I > NO3, Br > CI > acetate | > SO4 > tartrate | > citrate. The trivalent citrate ion is a weaker swelling agent than the bivalent tartrate and sulphate ions and these last two are weaker agents than the monovalent ions. In the case of positively charged proteins with cationic polypeptide chains, Hofmeister's ion series appears to be reversed, because the adsorption now refers to the anions. This inversion is particularly striking if one succeeds in reversing the charge of a negative gel. For instance, with gelatin in a neutral or basic medium, where the gel IJZ FINE-STRUCTURE OF PROTOPLASM II framework acts as an anion, the order in which ions furtlier swelling is as follows Li > Na > K > Rb > Cs. In an acid medium, however, in which the gel framework behaves like a cation: Li < Na < K < Rb < Cs. Now one would expect that in the isoelectric, i.e., uncharged state, the gel frame would show the same degree of swelling in all neutral chlorides, since in that case no electrostatic attractive forces are operating. This is not so, however; one finds so-called fransitionarj series which are of special importance for biology : Li > Na > K < Rb < Cs. This result is not easily comprehensible after what has been said before. For, if one plots the degree of swelling against the atomic weight of the cations, one obtains a descending curve in alkaline solutions (gel framework negative) and an ascending curve in acid solutions (gel frame positively charged; Fig. loia). For this reason one would expect a horizontal hne if the pjj of the swelling medium has been adjusted to the isoelectric point (L E. P.) of the protein. However, the experiment yields a minimum curve in which K holds a special place. By using ion models, however, it is possible to understand these relations, too. It follows from the ion mobilities that the two ions in KCl are of equal size. For this reason they are adsorbed in the same way by an isoelectric framework. In LiCl and NaCl, however, Li and Na are adsorbed to a smaller extent than CI because of their large hydration layer. Consequently, the molecular framework assumes a weakly negative charge and is more strongly hydrated than in KCL Conversely, in RbCl and CsCl the cations are more easily adsorbed than the CI ions, which again results in a weak electric charge of the gel, accompanied by increased hydration (Fig. loi b). Since the isoelectric point of cytoplasm usually lies in the weakly acid region, cations have a discharging effect on it. As a rule, therefore, cations cause less swelling than water. Anions, on the contrary, in- crease hydration as a result of their similar electric charge, so that cytoplasm often swells considerably (cap-plasmolysis, p. 197) in par- CYTOPLASM 155 ticular with SCN, I, Br, but also with CI, which is held to be re- sponsible for the swelling of the cytoplasm of haloph^nies (Stocker, 1928). In the series I,Br,Cl,F, chlorine often takes a similar optimal place to potassium in the alkaline metals and Ca in the alkaline earths (PiRSCHLE, 1930). 1 1 1 ! Ooj Q. 1 i ' — Swelling m/nimum 1 1 1 Li No K oj Rb Cs Li No K Rb Cs b) Fig. loi. Change in swelling of a gel frame consisting of polypeptide chains under the influence of chlorides of the alkali series at various pn values ; Q = degree of swelling, a) Hofmeister's series; dotted line = behaviour expected at pn = I.E. P.; instead, one finds b) transitionary series. III. Heteropolar valency bonds. If all acid and basic groups in the cytoplasm exactly cancel each other out, the isoelectric state (im- properly called "isoelectric point" I.E. P.) is attained and nearly all properties of protoplasm reach extreme values : the degree of swelling becomes a minimum, the danger of setting a maximum ; the stability is low, the electric charge and the migration in an electric field become zero by definition, etc. If then positive and negative end groups of the side chains occupy suitable relative positions, they can enter into salt-like bonds (Fig. 96 III, p. 145). Their electric charges are neutralized and the hydration of the region in question is reduced to a minimum. The salt bonds cannot be broken as readily by neutral salts as the heteropolar cohesive bonds. Something more drastic is required, viz., the concentration of the H ions (pn), must be changed. Some of the intermolecular salt bridges are then hydrolyzed and a certain number of the bound carboxyl and amino groups become free. If hydrolysis is achieved by H ions, i.e., if the p^ of the liquid in which the cytoplasm is examined drops below the I.E. P., the dissociation of the free COOH groups is diminished, that of the amino groups (-NH3OH) is increased. Thus the cytoplasm acquires a positive electric charge and behaves like a complex cation. Conversely, if the p^ of the medium is greater than the I.E. P. of the cytoplasm, the dissociation of the COOH groups is ^54 FINE-STRUCTURE OF PROTOPLASM II increased and the cytoplasm becomes negatively charged, i.e., it acts like a weak anion. This occurs as a rule in neutral nutrients, since the I.E. P. of protoplasm is usually lower than 7 (Table XVII). TABLE XVII ISOELECTRIC POINT (i.E.P.) OF CERTAIN PROTOPLASTS (according to Pfeiffer, 1929) PH Bacteria: Fungi : Algae: Angiosperms : Bacteriutn coli .... grampositive bacteria gramnegative bacteria Fusarium Nitella Hyacinthus (root tip) . Lupinus, Pisum . . . Rheum Solarium 12-13 about 5 2-3 5-4 4.4-9.6 4-3 4-3 4.5-4.8 6.4 The isoelectric state determines the acidity at which the heteropolar junctions of the salt bonds are most effective. Any deviation of the p^ from this state results in a loosening of this type of bond. Up to a certain point esterifications, i.e., bridges formed between alcoholic OH and acid groups of neighbouring polypeptide chains (Fig. 96III, p. 145), can Hkewise be reckoned among the heteropolar valency bonds. Their firmness is dependent also on the pjj of the medium, since hydrogen ions are capable of hydrolyzing and hydroxyl ions of saponifying these ester bonds catalytically. IV. Homopolar valency bonds are formed either by elimination of water (ether, glucoside and peptide bridges. Fig. 96 IV, p. 145) or by splitting off hydrogen, i.e., dehydrogenation (methylene and sulphur bridges. Fig. 102). The former still possess a certain polarity and can be hydrolyzed under suitable conditions. Without the aid of enzymes this can now only be effected at temperatures above the physiological; compare, for instance, the hydrolysis of glucosides and proteins by boiling acids. This is of particular importance for the stability of the carbohydrates and the peptide bonds. The purely homopolar valency bridges (-CH2-CH2- -S-S-) can no longer be hydrolyzed at all. Here I CYTOPLASM 155 the loosening of the junctions is achieved according to an entirely different principle, namely by addition of elementary hydrogen ihydrogenation). cystine bridge: CH-R-SH HS-R-CH <— — ^ CH-R-S-S-R-CH / methylene bridge: CH-R-CH3 CH3-R-CH i===^ CH-R-CH2-CH2-R-CH -2H TTh Fig. 102. Bridges dependent on r^ At physiological temperatures water in very small amounts not only dissociates into ions according to the scheme HgO ^ H+ + OH~, but also, though admittedly to still less extent, into the elements hydrogen and oxygen: zHgO^ 2H2 ^ Og. These gases develop a very low gas pressure, which for hydrogen we shall designate as tHg. If the partial pressure of hydrogen in the cytoplasm increases, the -S-S- bridges tend to be hydrogenated, which causes rupture of the bonds (Fig. 102). The cystine bridges can therefore absorb Hg and for this reason act in the same way with respect to the partial pressure of Ha as a buffer with respect to the concentration of H+ions. These conditions have been investigated in particular in the case of gluta- thione (G). This is a protein compound which can be split into glutamine, cysteine and glycine. It represents a tripeptide chain with the three amino acids mentioned as side chains. However, whether it occurs in the cytoplasm as a free molecule or only as part of a much larger macromolecule cannot be decided at present. In both cases glutathione reacts according to the following scheme : 2 GSH ^ GS-SG + Hg. Thus, when sulphydryl groups occur in the side chains of proteins (Fig. 96 IV, p. 145), these can give rise to formation or dissolution of cross-links. Methylene bridges cannot be formed with the same ease, at any rate in the laboratory, where methyl groups show a very passive behaviour. All the same, it is known that in the metabolism one molecule of succinic acid can be formed out of two molecules of acetic acid by dehydrogenation (Mothes, 1933). This succinic acid is then dehydrogenated further to fumaric acid, converted into malic acid, dehydrogenated to oxalo-acetic acid and finally, after decarboxylating this keto-acid, converted into pyruvic acid. It thus becomes apparent 156 FINE-STRUCTURE OF PROTOPLASM II that dehydrogenation plays an important part in the chemistry of fermentation. It is, therefore, likely that to a certain extent this also applies to the formation of methylene bridges between neighbouring polypeptide chains. It is known that in asphyxia the cytoplasm often liquefies; this may be due partly to hydrogenation processes, resulting from increased partial pressure of hydrogen. The hydrogen pressure in protoplasm is characterized by its negative logarithm in much the same way as the hydrogen ion concentration. The Ph is derived from the product of the ionic concentrations (cH+) • (cOR-) = 1 0-1*. Similarly, the product of the Hg and Og partial pressures in water is constant. It amounts to (tH2)'^-t02 = iQ-^^ in which the pressures are expressed in atmospheres. Thus the Hg and O2 pressures are mutually dependent in the same way as the H+ and OH" concentrations. The hydro- gen and oxygen pressure or the so-called redox potential of a solution in water can therefore be characterized by a single number. For this purpose we choose the negative logarithm of the hydrogen pressure, which is designated as r^. If hydrogen is made to flow through a system under atmospheric pressure, the hydrogen pressure amounts to i atm., or, written in exponential form : 10" atm., which means that r^ = o. On the other hand, if oxygen flows through the system, tOg = i, and accordingly (tHg)" = iQ-^^ or ry = 41. Obviously the rjj of a system can vary between o and 41. Small values of r^ indicate lack of oxygen, larger ones on the contrary are indicative of fav- ourable aerobic conditions, r^ (like p^) can be measured directly with the aid of a potentiometer (Fig. 103) or indirectly with the help of suitable dyes (MiCHAELis, 1933) which lose colour at a certain r^ as a result of hydrogena- tion (for instance, methylene blue and indigo). The analogies between p^ and Th are listed in Table XVIII. The characteristic values of the rjj scale are apparent from the following list : % I at. Og (oxygen electrode) 41 air (1/5 at. oxygen) 4°^1 hydrogen and oxygen pressure in equilibrium .... 27.3 H2 pressure =2-02 pressure (middle of redox scale) . 20.5 , 1 ^ ( aerobic life o border or ; 1 ■ i-r ° ( anaerobic lire I at. H2 (hydrogen electrode) o Table XIX gives a few r^ measurements in living cytoplasm (Needham, 1925, RiES, 1938). The values are not strictly comparable, since according to the equation Hg — 2 el~ ^ 2H+ the value of rjj is a function of p^- This dependence is apparent from Fig. 103 (according to Bladergroen, 1945). If the electric redox potential E (with respect to the platinum hydrogen CYTOPLASM 157 electrode E = o) and the value of p^ in the system are used as rectangular coordinates, the curves of constant hydrogen pressure (rjj) are sloping lines. If two of the three quantities: electric redox potential E, the exponent of hydrogen pressure r^ and the exponent of hydrogen ion concentration pjj, are known, the magnitude of the third one can be read from the diagram in Fig. 103. Since the redox system is only determined by its electric poten- tial E, it follows that in biological systems both the t^ value and its corre- sponding Pjj value should be given. On this condition the rj^ value may be identified with the redox potential, as is usually done in biology. TABLE XVIII Pjj and rjj SYMBOLISM Actual acidity Ph Redox system ■■H Starting point. . . Dissociation . . . Law of mass action Exponent Inter\^al hydrogen ion cone. cH+ H2O ^ H+ + OH- cH+-cOH- = 1 0-1* Ph = -log cH+ Pjj varies from 0-14 hydrogen pressure tHg 2H2O ^ 2H2 + O2 (tH2)2-t02 = IO-«2 r„ = -log tHo ■■H rjj varies from 0-41 TABLE XIX REDOX POTENTIAL (rjj) OF CERTAIN PROTOPLASTS (according TO NEEDHAM, 1 92 5) Sea-urchin egg Amoeha proteus Salivary gland of Chironomus . Ph 6.5 7.6 7-2 ^H 19-21 17-19 19-20 Just as the heteropolar valency bonds are strongest at a certain pjj, namely at the I.E. P., there is an optimum value of Tjj at which the homopolar valency bonds are the least endangered. It has already been pointed out that cystine bridges are broken down at high hydrogen pressures, i.e., at low rjj-values. At high values of rjj they are re- established. A high rjj is, however, also capable of loosening bonds (oxidation). As shown by Staudinger (1957a, p. 13), the glucoside 158 FINE-STRUCTURE OF PROTOPLASM II oxygen bridges of cellulose from a certain degree of polymerization onwards are very sensitive to oxidation, so that the chains are easily degraded, for instance according to the scheme: (C6Hio05)2n -f O., = 2(QHio05),0. Similar sensitive ether bridges may be assumed to exist in the cytoplasm, so that not only too small a r^ but also too high a r^ may interfere with the bonds inside a macromolecule. Apart from dehydrogenation, i.e., elimination of hydrogen, the transfer of hydrogen atoms from one chain to a neighbouring chain may also be responsible for bridge formation. Astbury (1936) and AsTBURY and Wrinch (1937) dis- cuss two possibilities of bridge form -ation inside folded polypeptide chains of fibre proteins, and similar reactions may also be considered in protoplasm. The hydrogen can be exchanged between neighbouring keto and imido groups following the lactam-lactim tautomerism ac- cording to the abbreviated equation > CO + HN< ^ >C(OH)-N< , thus building a main valency bridge. In the same way bridges may be formed between keto and methylene groups by keto-enol in- version: > CO ^ RHC < ^ > C(OH)-RC < . Such inversions Fig. 103. Relation between redox po- tential E, hydrogen ion exponent pjj and hydrogen pressure exponent r^. Abscissa : PH-value of the system; ordinate: electric potential E volt of the system with respect to Pt-Hj-electrode (from Bladergroen, 1943)- often occur quite easily, and in many cases it is impossible to decide which of the two forms is present. In the case of cytoplasm, this would mean that because of the possibilities discussed it would remain doubtful whether a bridge existed or not, i.e., its existence might be obvious at one moment and fail at the next, which would be in accordance with the great instability of the bond and with the mobility of the cytoplasm. V. Long-range forces . Whereas the forces described under I-IV have an extremely small radius of action, there are reactions between I CYTOPLASM 159 protein macromolecules, submicroscopic and even microscopic pro- tein particles which bridge submicroscopic distances. Such reactions occur when rod-shaped virus particles (Fig. 84c, p. 126) take a parallel orientation in a concentrated sol (Wyckoff, i947a-c), when protein macromolecules aggregate according to the rule of Svedberg (Fig. 84b) or when globular submicroscopic particles crystallize (Fig. 84 d). Similar attractions over considerable distances appear when antibodies (precipitins, agglutinins) cause the precipitation of specific proteins or even the agglutination of bacteria and blood corpuscles. The nature of long-range forces is difficult to understand. As their radius of action is greater than 50 A, they play an important role in the structural arrangement of colloidal particles. Oster (195 i) shows that long-range orientation is partly due to the repulsive effect of electrical double layers in highly concentrated sols, and partly to ordinary Van der Waals attractive forces which are additive, so that an integrating effect of all the atoms of two adjacent macromolecules is involved. ROTHEN (1947) has published experiments indicating that the action of long-range forces is detectable at distances of over 200 A. He coated the antigen of bovine albumin on a slide with a layer 200 A thick of barium stearate and was able to observ^e the immunological reaction of the antibody applied to this film. Even enzymes such as trypsin and pepsin were found to act upon substrate layers through an inert screen. The last experiment is in contradiction to the current conception of enzyme action, which is considered to be a contact reaction with the molecules of the substratum. The impermeability of the intervening stearate films has therefore been doubted (Trurnit, 1950). Whatever the result of this criticism may be, long-range forces incontestibly cause the aggregation of submicroscopic particles in sols and the formation of structures in gels. There must be a discrete number of spots on the surface of a globular macromolecule where junctions are possible. If this number is two, the protein globules have a tendency to form beaded chains (Fig. 104), which may yield a loose reticulum. If the number of active spots is three, they will be the origin of a two-dimensional layer repre- senting a porous film (Fig. 104). Four junctions would cause a three- dimensional framework, since they are arranged rather in a tetrahedral manner than in a plane. A sphere may touch as many as 6 neighbours [6o FINE-STRUCTURE OF PROTOPLASM II in a plane. This yields a dense film. When several layers of the kind are superposed, a close-packed crystal lattice results (Fig. 84d, p. 126); in this case every macromolecule is fixed by 1 2 junctions. This suggests that the junctions are induced wherever the globules touch. Although this seems true for crystallizing proteins, it would be difficult to under- O Fig. 104. Aggregation of globular macromolecules (dots = spots of junctions), a) Two spots of junctions produce beaded chains; h) three spots of junctions produce porous layers; c) four spots of junctions produce tetrahedral groups; d) twelve spots of junctions produce a close-packed crystal lattice. stand the formation of beaded chains and loose meshworks without assuming a restricted possible number of junctions per aggregating particle. In the case of globules aggregating to beaded chains, the macromolecules must be endowed with a pronounced polarity. Summary. The proteins are to be considered as the structural ele- ments of the cytoplasm. Their macromolecules are interlinked to form a framework, whose junctions can be disrupted by various quite different agents. A rise in temperature attacks in the first place the homopolar cohesive bonds or lipidic bonds. Dependent on their state of hydration, adsorbed salts affect the heteropolar cohesive bonds or I CYTOPLASM l6l secondary valency bonds; p^ influences the heteropolar or salt bonds, and the redox potential is capable of intervening, either as a con- structive or as a destructive factor, in the homopolar valency bonds or bridgelike bonds. It is therefore very difficult to explore the structure of the cytoplasm experimentally, for it is scarcely possible to vary only a single one amongst these four factors, keeping the three others rigorously constant. A change in the temperature or the salt concen- tration will often cause changes in pn and tn, which in their turn are interdependent. For this reason one can never be sure in an experiment whether some measure has not affected other tj^pes of bonds besides the group of points of attachment which one wished to investigate. In spite of the fact that it is practically impossible to keep the four types of junctions as neatly apart as in theory, this scheme gives an idea of the various kinds of bonds which by their harmonious col- laboration are responsible for the remarkable structure of protoplasm. In the case oi fixation, the aim is to preserve the molecular arrange- ment as true to life as possible. This can never be done perfectly, since the usual means of fixation affect quite different categories of junctions. Alcohol has a dehydrating and hardening effect on the heteropolar cohesive bonds. In order to counteract the accompanying shrinkage, a swelline medium such as acetic acid has to be added. Its H-ions lessen the contracting action of the alcohol by hydrolysis of hetero- polar valency bonds and by maintaining a certain state of hydration of the heteropolar cohesive bonds. Oxidizing fixatives like chromic acid and osmic acid affect bridges which are sensitive to r^^ and thus solidify the labile hompolar main valency bonds. The tanning action of formal- dehyde may be due to its capacity to form bridges between neighbour- ing polypeptide chains according to the same scheme as that which governs the polymerization of oxymethylene. It is impossible to find a fixation mixture which in no way affects the structure of the labile cytoplasm. In spite of this, fixations which have been carried out correctly cannot be compared with precipitations, since there is no separation of phases, but only a coarsening of an existing structure. It is shown by the dyeing experiments carried out by Drawert (1957) with varying p^ of the imbibing liquid, that the molecular framework after fixation still contains acidic and basic groups capable of dis- sociation, although these groups are no longer screened off but are freely accessible to dyes. This is why fixed cells can be stained easily, whereas ;6z FINE-STRUCTURE OF PROTOPLASM IE vital staining of living cytoplasm is almost impossible. The enhanced adsorbing power of dead cytoplasm allows of identifying dead cells with the fluorochrome acridin-orange (Strugger, 1949)- Depending on its concentration, this dye shows a green (1:50,000) or a red (1:100) fluorescence in the UV light. Since dead cytoplasm adsorbs a considerable amount of acridin-orange, it displays a magnificent red fluorescence, whereas living cells appear to be green. With the aid of the diagram of Fig. 96 (p. 145) some indication of the sio-nificance of the various elements in the structure of the proto- plasm can be given. In the periodic system (Table XX) all elements which are of importance to the life of plants lie on a line connecting carbon with the inert gas argon. I have designated this line as the nutrition line (1935c); only hydrogen and molybdenum (Arnon and Stout, 1939) are an exception. TABLE XX elements which are indispensable to plant nutrition Series 0 1 Q IQ H V B ra m. 0 Is* penod H He 2nd penod He U Be B A^ N. 0^ s F Ne 3rd pcnod Ne Nal ."» -^ « Is. P a ~\ ~ ^. iHi (»nod (A/ 0. (j Zn Sc r, 6e V As Cr 5e Mn Br ^ Co Ni Kr 5 Hi period Kr Rb Sr Cd Y In Sn Nb Sb Mo Te Va Ru Rh Pd X Gtti penod X Oi Ba •n Ce Pb la Bi W ft) Re Os Ir P» Rn 7 Hi period Rn - Ra fit T>. ft U In Table XX the indispensable elements have been framed by squares, whereas those which are found in nearly all plants, but whose indispensability remains to be proved, have been framed in dotted lines. C and N Ue in the centre. These elements occupy a central position in the molecular structure, too, since they form the poly- peptide main chains. They may therefore be designated as chain- building elements. The chains are built according to the scheme -C-C-N-C-C-N-. Notwithstanding its close relation to nitrogen, phosphorus does not occur as a chain-building element in this manner, but only in combination with oxygen (compare Fig. 122, p. 215: -C-0-P-0-C-) ; as in the inorganic domain, it is always present in an oxidized form as phosphoric acid. In the degradation of carbo- I CYTOPLASM 163 hydrates it also acts as a protector of atom groups which should not be affected (hexose diphosphoric acid, phosphorus glyceric acid, etc.). It is possible that in the cytoplasm the phosphatides, which can combine with various groups of the polypeptide chains, render a similar service. The elements O and S of the sixth row are primarily bridge-building elewenis, since they interconnect the C-N-polypeptide chains. Apart from this, oxygen can act as a chain-building element in the high polymeric carbohydrates, and conversely N and C are capable of bridge formation. The elements of the first and second row: Na, K, Cu, Mg, Ca, Zn, and also CI occur in cytoplasm as ions and act as hydration regulators. They do not form stable bonds but only heteropolar salt bonds with the molecular framework (metallic organic compounds like chloro- phyll, haemoglobin, etc. are quantitatively of minor importance). In this respect the most favourable ions in plants are K, Ca and CI of the so-called argon type (in animals Na takes the place of K). Both in mixtures and in pure solutions these ions are tolerated in concen- trations at which other ions are detrimental to the cytoplasm structure. This would also explain why the nutrition line takes its course towards argon. The higher valent elements B, Mn and Fe presumably enter into some relation with the protoplasmic frame. As regards manganese and iron, it is usually believed that their capacity to change valency is put to use in metabolism. The most important part is played by the element hydrogen, both as an ion and as an element. It regulates p^ and ry, thus preventing the molecular framework from soUdifying, and maintaining the labile changeable state which is so characteristic of protoplasm. c. Physical Properties of the Cytoplasm Sol properties. Many cytologists suppose the cytoplasm to be a Hquid (Rhumbler, 1898). Heilbrunn (1930), for example, writes about the amoeba: "it is a tiny sac of fluid in motion" and Chambers (1925) considers not only the cytoplasm but also the nucleus to be a liquid phase. The flow of protoplasm, the relatively low viscosity, the large water content, the soft consistency, the convex shape in plasmolysis and other indications point to the sol character of the cytoplasm, i.e., to a state in which all submicroscopic particles have free relative move- 164 FINE-STRUCTURE OF PROTOPLASM II ment. The most striking of these effects \& protoplasmic flow, and when seen for the first time this phenomenon will always convince the observer of the liquid state of the cytoplasm. The merit of having characterized the aggregate state of cytoplasm with the aid of physical laws is due to Rhumbler (1914)- According to his observations, the cytoplasm of the amoeba possesses i. no measurable elasticity, 2. no perceptible compressibility at ordinary pressures and 3. it follows the capillary laws which are determined by the surface tension (minimum surface, constant contact angle, spread- ing on the surface of liquids, capillary rise). At the present time our picture will be somewhat different. According to Newton's law, ideal Hquids are completely free from inner elasticity: any particle in the bulk of 'the liquid can be moved at will without showing the slightest tendency to swing back into its original position. In cytoplasm this condition is not fulfilled, for, as will be shown below, it possesses structural elasticity or elasticity of flow. The incompressibility should not be tested at "ordinary" pressures, but at high pressures where the low compressibility in comparison with solid bodies becomes apparent. If a living amoeba in its nutrient is exposed to a uniform pressure of the order of magnitude required to prove incompressibility, its cytoplasm is altered, whereas it is the main property of ideal liquids not to undergo any changes in this experiment. Brown (1934) and Marsland (1942) show that the cyto- plasm of different eggs, of Amoeba, Paramaecium, of human erythro- cytes and of Elodea leaves becomes liquefied by high hydrostatic pressure. It behaves therefore Uke sols in which the process of gelation is accompanied by a small increase of volume. According to obser- vation in the centrifuge microscope with a high pressure chamber, the mobility of included particles increases by almost 25% for each pressure increment of 70 atm. (1000 lbs/in.^). Under these conditions protoplasmic streaming is inhibited, and within fairly broad limits, the effect is reversible. Pressure up to 300 atm. may be maintained for about an hour, and yet, when the cells are returned to atmospheric pressure, the original structural characteristics are restored within a minute. At 700-1000 atm. even the cortical layer of the cytoplasm is liquefied and irreversible changes begin to appear. Rhumbler's best arguments refer to the capillary properties of naked cytoplasm, although by no means all cytoplasts assume a I CYTOPLASM 165 Spherical shape or can be spread at will on the surface of another liquid. In those cases where the cytoplasm forms liquid drops, its surface tension y can be measured (E. N. Harvey, 1936) by observing the cell as a sessile drop flattened by gravity. The relation y == g (d - d')r-^F is used to calculate y; g is the acceleration due to gravity, (d — d') the difference in density between drop and medium, r the radius of greatest flattening and F a function of the distance a in Fig. 105 a representing the flattening of the drop. For the egg of the mollusc Busjcon canaliculatum a tension of 0.5 dyne/cm is found by this method, while the egg of the salamander TriturHS virescens gives only o.i dyne/cm (Table XXI). The eggs of mackerel contain a large oil droplet which can be flattened against the rigid cell membrane when revolving the egg at high speed in the centrifuge microscope of E. N. Harvey. From its shape, an oil/cytoplasm interfacial tension of 0.6 dyne/cm is calculated: if the centrifugal force is increased up to 450 times gravity, this tension does not change, showing that the surface is not elastic. In contact with sea water this oil gives a tension of 7 dyne/cm, a high value which is explained by the rule that the interfacial tension be- tween two immiscible liquids is the difference of the tensions of the two liquids against air. As the surface tension y of water is 72 dyne/cm and that of oils is only about half as much, it is evident that the cell surface cannot be formed of pure lipids, because this would provoke a higher interfacial tension between the surface of a cell and its culture medium. A surface with only o.i dyne/cm tension against the medium cannot be very lipidic; besides the lipids it must contain proteins with a certain affinity for water. If the cell does not flatten under its own weight, the flattening can be achieved by compression (E. N. Harvey, 1937): the spherical cell is loaded by a thin beam of gold with micro weights. The weight W divided by the area D of the flattened cell in contact with the beam gives the pressure P, from which the surface tension is calculated by the formula l66 FINE-STRUCTURE OF PROTOPLASM II IV oT when Tj and fg are the two radii ofthe flattened cell indicated in Fig. 105b. The unfertilized egg of the st2i-\xtchin, A.rbacia punctnlata, shows a surface tension of 0.135 dyne/cm when loaded with two micrograms. Smaller weights give lower values and extrapolation of the tension/ compression curve yields 0.08 dyne/cm for the uncompressed egg. As the surface tension is not con- stant but depends on the interior pressure, the surface displays elasticity : this again is evidence of the presence of proteins in the cytoplasm surface, since a layer of pure lipid would not show elasticity. Sols have no elastic properties, so it is evident that the proteins in the surface layer are in a gel-like state. Fig. 105. Measurement of surface tension (after E. N. Harvey, 1936/37); a) sessile drop, b) flattened drop. TABLE XXI SURFACE TENSION OF PROTOPLASM WITH RESPECT TO SOLUTIONS (according to E. N. HARVEY, 1 93 7) Naked protoplasts Leucocytes (Lepus caniculus) .... „ {Rana pi pi ens) Amoeba {Amoeba dnhia) Slime mould {Physarum polycephalum) Sea-urchin egg {Arbacia punctulatd) . Salamander egg ij'riturus viridescens) Medium Ringer sol. + serum 5? H " „ , diluted „ , 250 X diluted seawater pond- water +gum arable It is clear that the occurrence of capillary phenomena gives no conclusive evidence of the existence of a true liquid. On the other hand, however, it has not been proved that Hquid cytoplasm possesses an organized structure; it has only been shown that the possibility of such a structure cannot be excluded. The same holds good for the results of viscosity studies on liquid cytoplasm, which give valuable information on changes in fluidity. 1 CYTOPLASM 167 Viscosity measurements can be performed by examining the Brownian movement of granule inclusions (Pekarek, 1930) or by observing the speed of a heavy particle falling through the cytoplasm by its own weight, or by centrifugal force (Heilbronn, 1914; Heilbrunn, 1930). The intensity of Brownian movement is given by X2 RT I t N ^Ttrrj where X^ represents the mean square of the displacement of a granule with radius r during time t, Ris the gas constant, T the absolute tempera- ture, and N is Loschmitt's number. It is seen that the viscosity r} of the medium is inversely proportional to the intensity of Brownian movement. For the movement of a particle through a Uquid (Fig. 1 12a, p. 192), Stokes' law 2ng (d — d'y T] = 9v holds good. Here v is the velocity of the moving spherical particle, (d — d') the difference in density between cytoplasm and observed particle, g the acceleration due to gravity, and n the number of times which the applied centrifugal force is stronger than gravity. With these methods it is found (Table XXII, p. 1 69) that the sap in the vacuole of plant cells is often only about twice as viscous as water (Frey, 1926c). For the cytoplasm, however, relative viscosities of six in Amoeba (Pekarek, 1930), twenty-four in parenchyma cells of the Viciafaba seedling (Heilbronn, 19 14) or thirty in erythrocytes of man (Ponder, 1934) are found. Such values are more reliable if derived from Brownian movement than if determined by Stokes' law, since the latter requires a uniform velocity v of the faUing particles which is not often realized in cytoplasm. Once again these measurements do not establish the existence of structural viscosity in cytoplasm. To solve this question it is necessary to carry out viscosity measurements with different pressure gradients. Since protoplasm cannot be made to flow through a narrow tube like a liquid, Pfeiffer (1936) sucks naked protoplasts (so-called gymno- plasts from the decomposing fruit pulp of Pbysalis, Solanum or Juni- perus, of Allium epidermic cells, etc.) through a capillary under a given ;68 FINE-STRUCTURE OF PROTOPLASM II „}00 ^ 30 70 SO ^ ^^^ E \ V \ X V / a Pressure p in cm H2O pressure difference which can be read from a manometer. At the same time he measures the viscosity by following the Brownian movement of particles (dyed by means of chrysoidin) which are embedded in the protoplasm (Pekarek, 1932). In Fig. 106 the viscosity 7] is plotted against the pressure gradient p for plasmic drops from the cells oiCbara fragiUs. The viscosity decreases rapidly with increasing pressure (measured in cm HgO), where- as in normal flow of glycerin r\ remains practically independent of the pressure. This experiment shows clearly that protoplasm is not a sol-like liquid, but repre- sents an elastic^' gt\ solution". This does not yet imply a definite struc- ture, although once more this pos- sibility is not ruled out. It is otherwise with the devia- tions from Stokes' law. Ac- cording to this law, microscopic- ally visible particles or bubbles in a liquid either fall or rise with constant velocity. Scarth (1927) has ascertained, however, that in cytoplasm the particles do not move with uniform velocity. It looks as though they encounter invisible obstacles, and they fall in a hesitant and jerky manner. According to Scarth, they give the appearance of lead shot which is run through a brush heap. Again and again the falling particles meet with invisible strands, lose speed and change their direction. Accordingly, the cytoplasm cannot be homogeneous but must be full of invisible fibres of a higher density. It does not possess a uniform viscosity, and the results derived from the fall method (Heilbronn, 1914) represent some kind of average value. In Pekarek's viscosity measurements (1930, 1952), which are based on the amplitude of oscillation of particles in Brown- ian movement, the inhomogeneity of the cytoplasm is less apparent, because the oscillatory motion daes not cover a long distance through the cytoplasm and can be studied at a fixed point. The values reported for the relative viscosity of the cytoplasm do not prove its true liquid state, even though they are considerably Fig. 106. Structural viscosity of the cyto- plasm of Char a fragilis (from Pfeiffer, 1936). Abscissa: pressure p in cm HjO. I Cytoplasm at 21° C, II at 12° C; III glyce- rol at 21° C. Ordinate: Viscosity rj in % of the original value. I CYTOPLASM 169 lower than the values for many true viscous liquids (Table XXII). For a true liquid should in the first place be homogeneous in the physical sense and this certainly does not apply to cytoplasm. The following comparison may be permitted : Consider a wad of thread-like algae. The threads can be moved at TABLE XXII RELATIVE VISCOSITY fj Water * i Cel/ sap.- Stem parenchyma, oi the Vida Faba seedlinp: ... 1.9 (Heilbronn, 1914) Protonemz of Lepiobrjum piriforwe 1.9 (Pekarek, 1933) Epidermic cells of the .^///«w (7f/>a bulb 2 (Pekarek, 1930) Terminal vacuole of C/i9.r/cr///w (see Fig. 1 1 2a, p. 192) 2.5 (Frey, 1926c) Cytoplasm : Amoeba 6 (Pekarek, 1930) Stem parenchyma of the Vicia Faba^eeAhne. ... 24 (Heilbronn, 19 14) Red cell of man 30 (Ponder, 1934, p. 87) Viscous liquids: Glycerol 87 (L.a.ndolt-B6rnstein, 1923) Paraffin oil 92 ,, Castor oil 1250 ,, wiU with respect to each other, although they impede each other's freedom of movements as a result of their extremely anisodiametric shape. When transferring this microscopic model to the molecular domain, the threads become long chain molecules in a dispersing medium and a drop of this macromolecular sol would show structural viscosity and all the capillary phenomena described. If the individu- alized algae threads of our model were replaced by the graceful reticular alga Hydrodictyon (Oltmanns, 1922, p. 277), scarcely any change in the inner mobility of such a wad of algae would be observed. On a molecular scale this means that a drop which contains a coherent three-dimensional molecular network, instead of free chain molecules, will not only assume a spherical shape but also show a constant contact angle and spread on the surface of suitable liquids. In spite of this, the structural elements of the network cannot move freely! The network is so flexible, however, that its shape within the drop is determined by the forces of surface tension. /Ul the same, we cannot lyo FINE-STRUCTURE OF PROTOPLASM II Speak of a true liquid, for, when static equilibrium is established, the drop is inhomogeneoiis, not only at the surface, but also in bulk. To sum up, it can be said that cytoplasm in its Hquid state obeys neither the laws of Newton (Pfeiffer, 1937) nor those of Poiseuille or Stokes (Frey-Wyssling, 1940 a). Although to cytologists it may have the appearance of a liquid, it certainly is no true liquid in the physical sense. We had better not attach too much value to this simi- larity, for we should then be unable to penetrate its submicroscopic fine-structure, since a liquid possesses a structure only in its surface. On the contrary, it is my aim to stress especially the deviations from the physical laws of liquids, as it is precisely these deviations which offer us the chance of elucidating the structural properties of cyto- plasm. Gel properties. Often cytoplasm does not flow in Hquid drops, but shows plastic properties. This in itself would not be sufficient to indicate a solid state ; but it is also elastic and to a certain extent pos- sesses a constant shape. The result of plasmolysis is not always separ- ation from the cell wall of a definitely convex drop. On rapid de- hydration with strongly hypertonic solutions the shape in plasmolysis becomes concave or angular, indicating a certain rigidity of the cyto- plasm in this state (Prudhomme van Reine, 1955). Especially interesting is the spinning capacity of the cytoplasm, which is apparent from the fact that long strands can be drawn from it (Seifriz and Plowe, 193 i). Often this phenomenon can also be ob- served during plasmolysis in the form of the so-called strands of Hecht (Fig. 107a), although this name is scarcely justified, since their importance was pointed out by Chodat (1907) many years before Hecht (191 2). From Fig. 107 a it is apparent that spherical boundaries as claimed by Rhumbler (1898) occur only in a few fibres in a very imperfect form. A similar fact, which shows the non-liquid state of the cytoplasm, is the "angular plasmolysis" of sea-urchin eggs (Runn- STROM and Monne, 1945 ; Runnstrom, Monne and Wicklund, 1946). In the plasmoptysis of Spirog^ira cells the protoplasm can be drawn out into a long strand which contracts rhythmically (Fig. 107 b). Seifriz (1929) has shown that the cytoplasm of amphibian red cells can be drawn out to three times its normal length and the nucleus even up to 20 times its original length without the occurrence of any drops. All these properties of the cytoplasm are inconsistent with the CYTOPLASM 171 hypothesis of a true liquid; they point rather to some fibrous sub- microscopic structural element. The inner elasticity can be demonstrated by suspending iron filings in the cytoplasm and moving them by means of a magnetic field. As soon as the field is switched off, the particles swing back elastically to Fig. 107. a) Plasmic strands of epidermic cells from the bulb of Allium, plasmolyzed by CaClj, (according to Kuster, 193 5^) ; h) plasmic strand oiSpirogyra, extruded in plasmoptysis (from Frey-Wyssling, 1940^). their original positions (compare Heilbronn, 1922). This method has been further developed by Crick and Hughes (1950) to measure the internal elasticity of cytoplasm quantitatively. They find the modulus of rigidity of chick fibroblasts in tissue culture to be of the magnitude of 100 dynes/cm^. At the same time they give evidence of the thixotropic behaviour of the cytoplasm which can change its state reversibly from solid to fluid when stirred. In this respect it gives similar results to elastic gels of sodium oleate or bentonite. The reversible gel-sol transition is one of the most important proper- ties of cytoplasm, as it is the basic phenomenon in protoplasmic flow l-jZ FINE-STRUCTURE OF PROTOPLASM II (p. 1 86). If a gel is liquefied under isothermal conditions, the volume can either increase (gelatin, agar) or remain constant (Na-oleate and other thixotropic gels) or decrease (methyl cellulose in water), A decrease in temperature or an increase in pressure favours gel form- ation in the first case and sol formation in the third (Freundlich, 1937). Cytoplasm belongs to the third category (p. 187).^ In addition and in contrast to all other gels, it can also change its aggregate state by itself, even if the external physical conditions remain unaltered. It seems that anaesthetized cytoplasm is more gelated than in the active state, as Seifriz (1950) finds that any anaesthetic causes a reversible gelation of protoplasm. We are thus faced with the paradox that cytoplasm simultaneously exhibits the characteristics of hquids (fluidity) and of solids (elasticity). It is noiv solid, then liquid to an extent rarely observed in any other colloid. The task of submicroscopic morphology consists, therefore, in drawing up a structural scheme which explains the double nature of cytoplasm at the boundary of the two classical aggregate states. By doing so, we should gain more than by adhering to the concept of cytoplasm either as a liquid or as a gel, neither of which can be true in a general sense. d. Submicroscopic Structure of Cytoplasm Particulate globules. If we disregard the microscopic inclusions in cytoplasm (plastids, mitochondria, lipid globules, granules etc.), it represents a microscopically homogeneous pseudophase. This is no longer true when it is observed in the electron microscope, where submicroscopic particles appear to be dispersed in a reticulate, fiorous or homogeneous matrix of diameters from 5 00 to 1 5 00 A (Claude, 1946; Faure-Fremiet, Bessis and Thaureaux, 1948; Lehmann, 1950). In liver cells these particles are distinctly smaller than the mito- chondria, which measure 2000 to 5000 A. Claude suggested calling them "microsomes". Globules of 1000 A diameter may lodge as many as 64 of the biggest macromolecules listed in Table XV (p. 141), so the microsomes must contain a great number of protein molecules and other compounds. According to Bensley (1943), the submicroscopic particles, isolated from homogenized liver tissue by the centrifuge, consist of protein, ^ Brown (1934) and Marslakd (1942) have checked this up to 1000 atm. I CYTOPLASM 173 nucleoprotein, flavoprotein, triglycerides, lecithin, sterine, vitamin A and 80-90% water. They contain the ribonucleic acid of the cytoplasm (Jeener, 1948). According to the view of Caspersson (1941), they are involved in protein synthesis. As metabolic centres they are analogous to the mitochondria or chondriosomes, which, however, are microscopic particles and repre- sent a special system in the cell which is designated as chondriome (GuiLLiERMOND, Mangenot and Plantefol, 1933; Bourne, 1945). The mitochondria of guinea pig liver tissue can be isolated (Hoerr, 1943) and analyzed. They are of lipidic nature (43.6%) but contain at the same dme two proteins of different I.E. P. They are free of lecithin and cephalin (Bensley and Hoerr, 1934). Faure-Fremiet (1946) gives for the same material somewhat different figures: Protein 64.6%, glycerides 28.8%, lecithin and cephalin 4.2%, cholesterol 2.25 %. At any rate there is no nucleic acid present. This is confirmed by the lack of UV absorption (Monne, 1948). According to Monne (1942 b), the mitochondria may be strongly hydrophilic. The rodlet shape of the so-called chondrioconts and the double refraction of the filamentous mitochondria from the intestinal cells of Ascaris megalocephala (Giroud, 1928) indicate an inner structure resembling a mesophase. Originally Bensley (1937) thought that the chondriosomes might be merely coacervates. Claude and Fullam's (1945) electron micrograms of fixed chondriosomes show a lipid cortex and a watery, less dense central zone. In addition Muhle- thaler, Muller and Zollinger (1950) have found that, in kidney cells, they are coated with a distinct submicroscopic membrane. In recent publications the mitochondria are considered as important bodies with special physiological functions (Claude, 1944), as certain enzymes are fixed on them. Hogeboom, Claude and Hotchkiss (1946) found cytochrome oxidase, and Leuthardt (1949) was able to localize the enzymes of the tricarboxyHc acid cycle on the liver mitochondria. Muller and Leuthardt (1950) and Brenner (1949) have demon- strated that the mitochondria of intact lymphocytes perform oxidation — reduction reactions. This means that the respiration is assigned to these bodies. The fact that the mitochondria are dispersed throughout the cytoplasm would account for continuing respiration of parts dis- sected from a living cell. It is probable that new mitochondria originate exclusively from 174 FINE-STRUCTURE OF PROTOPLASM II pre-existing mitochondria, similar to plastids, chromosomes and virus particles. Lehmann (1947) has proposed the term hiosomes for such bodies which are characterized by self-multiplication and endowed with specific functional tasks. Reticulate ground-cytoplasm. The matrix in which the microsomes and mitochondria are suspended has quite a different aspect, depending on the object under investigation and on the method of fixation used. Claude and Fullam (1946) speak of a fibrous ground texture in the cells of the guinea pig liver, Faure-Fremiet and co-workers (1948) of a reticulate ground-plasm in the amoebocytes of the snail. The cytoplasm of the thrombocytes in the blood is hyaline, alveolar or fibrous depending on the fixation with osmic acid, formalin or alcohol (Bessis and Bricka, 1948). Bretschneider (1950a) describes a three- dimensional network 400 A wide, partly beaded strands in the cyto- plasm of ciliates fixed with OSO4. It looks as though we are about to have a repetition of the cyto- logical discussions on the structures of fixed cytoplasm as seen in the ordinary microscope, this time with reference to the submicroscopic aspect. It is obvious that only the finest textures observed come any- where near the natural situation, while the coarser textures are only worth while considering in relation to a possible linear coagulation of previously filamentous submicroscopic structural elements. Rozsa and Wyckoff (1950/5 1) have found that the cytoplasm of the dividing cells in the onion root tip yields a beautiful dense reticulate structure with very fine meshes (smaller than 0.05 [x diameter) when fixed in neutral formaHn, whilst every acid fixative (especially OSO4 and acetic acid) furnishes a very coarse cytoplasmic reticulum with almost micro- scopic meshes (0.5 fi diameter). Bretschneider (1950c) has made a systematic study of the influence of fixation on the submicroscopic structure of cytoplasm as seen in the electron microscope, and has tested all the treatments used in cytology on the same subject (root tip of onion). The best fixation is obtained in Champy's and in Kopsch- Regaud's fluids (Fig. 108/1,2), which contain formalin and OSO4 combined with chromic acid and potassium bichromate. The hyalo- plasm shows a fine network of thin protein filaments with a diameter of about 160 A forming a regular hexagonal pattern. Pure solutions of formaHn (Fig. 108/4, 5), Bouin's fluid (Fig. 108/5) ^^^ Helly's fluid (Fig. 108/6) yield a slightly coarser network. Substances which CYTOPLASM 175 Fig. 108. Cytoplasm of the meristem cells of the root tip in onion. Pictures after different fixation fluids but at the same magnification of 12000 X and 100 kV. (By courtesy of L. H. Bretschneider, 1950c). coagulate the proteins strongly, such as acetic acid, trichloracetic acid, phosphotungstic acid, alcohol, sublimate or sulphosalicylic acid destroy the fine pattern of cytoplasm by syneresis. Contrary to Wyckoff's statements, osmic acid is found to produce fairly good fixation for animal cells. jj(, FINE-STRUCTURE OF PROTOPLASM II It is a remarkable fact that mixtures of fluids which fix different types of junctions seem to effect the best fixation (seep. i6i), whereas, with the exception of formalin, pure compounds produce poor fixation. A special feature of great importance is the occurrence of beaded chains observed in the electron microscope. Bessis and Bricka (1948) have described such microfibrils (of ^^ 500 A diameter) in the cyto- plasm of thrombocytes, Matoltsy, Gross and Grignolo (195 i) in the vitreous body of cattle eyes, and Lehmann (195 1) observes similar chains in the cytoplasm of Amoeba. Sheaves of such beaded chains with knots of 600 A diameter occur in liver cells; Bernhard, Gautier and Oberling (195 i) have shown that these beaded fibrils belong rather to the ergastoplasm subject to metabolic changes than to the mechanical cytoplasmic framework. In the egg of Tubifex, fibrils carrying knots of about 0.15 /i diameter have been found (Lehmann and Biss, 1949); these fibrils form the ground-plasm in which the microscopic yolk granules (2 ^t) are suspended. The knots (0.15/^) reach microscopic dimensions and are identical with the chromidia of Hertwig found in the sea-urchin and Tiibifex eggs. They contain ribonucleic acid (Monne, 1946a), The protoplasmic fibrils appear to be segmented by the chromidia and display for that reason a micro- scopical structure similar to the chromatids (see p. 225). Monne (1948) identifies these chromidia with the microsomes, be- cause both contain ribonucleic acid (Feulgen negative, UV absorption at A = 260 vafjL, stainable with pyronin), which differentiates them from the mitochondria. However, such an identification must be dis- carded from a morphological view, because the chromidia are im- movable bodies fixed on a beaded microfibril, whereas the microsomes are corpuscularly dispersed free and independent particles. In the gelated state cytoplasm has some continuous structure and, given the chemical composition of the cytoplasm (p. 140), it must be a protein gel. Protein molecules can aggregate to a framework in different ways. a. Globular molecules or composite submicroscopic particles may associate to form beaded chains (Fig. 5 la, p. 66). If these chains become sufficiently long or branch, a framework is easily formed. The gelation of gelatin belongs to this type (Joly, 1949). b. Expanded polypeptide chains can aggregate to form fibrils, such I CYTOPLASM 177 as are found in fibrous proteins, which may give rise to a meshwork or a plaitwork (Fig. 51b, p. 66). The first type of framework must produce gels with a higher percentage dry weight than the second, which we can picture as being made up of submicroscopic or amicroscopic strands. If the cytoplasm appears homogeneous in the electron microscope, the structural ele- ments (globules or threads) must be amicroscopic, i.e. they must have micromolecular diameters (< 50 A). It is difficult to decide which type is really present, because the structure easily changes in character owing to the denaturation of proteins in the fixation and drying processes. The inner structure of the globules and microfibrils is governed by the junction principles discussed on p. 145. If these sub- microscopic elements aggregate to form a gel, another type of junction is involved, caused by long-range forces (p. 158). The nature of these forces is not well known but in forming gels they act morphologically disjunctions in the submicroscopic domain in very much the same way as the chemical forces do in the amicroscopic range. According to OsTER (195 1) there is no real difference between short-range forces and long-range forces. Assuming that there is such a gel, all the cytoplasmic properties, strange as they may be, can be accounted for. The high water content of the cytoplasm (70 to 80% or more) is caused by the considerable width of the meshes of the framework. In addition, there is hydration water inside the submicroscopic strands and beaded chains. The water content is liable to be so great that many of the water dipoles are not fixed by the framework and have freedom of movement. In this case excretion of water from the cj^to- plasts and hence vacuolization becomes possible. As a rule, however, all the water is loosely bound by main chains or side chains and takes part in establishing the maximum state of swelling. The transition of protoplasm to a resting state is accompanied by a gradual diminution in the amount of water brought about by a narrowing of the submicroscopic interfibrillar and intramolecular interstitial meshes. The water is perhaps partly replaced by lipids, as hydrophilic groups are screened off by phosphatides, sterines and the like. The determinant structure and the organization of the framework which governs the processes of life can thus sometimes be preserved for years (spores, seeds). Evidently this natural deh3"dration cannot 178 FINE-STRUCTURE OF PROTOPLASM IE be imitated by artificial drying at room temperature, since the change- in the framework structure has to proceed step by step along with, the dehydration caused by the neutralization or screening of the hydrophilic groups, without changing those configurations of the molecular structure which are necessary for the maintenance of life^ But by the modern procedure oi free^e-drying a method has been found which permits evaporation of the hydration water without altering am~ structure essential to life. Freeze-dried bacteria can be preserved in- definitely ; and this method seems to be very promising for the preven- tion of denaturation when fixing submicroscopic protein structures^ The physical properties fluidity^ plasticity and elasticity must be at- tributed to the character of the junctions between submicroscopic particles. The more these are dissolved, the more liquid the cytoplasm, becomes. However, the junctions must never all be weakened at the same time. Jn other words, the cytoplasm must never become a true sol in which all particles can move freely. Certain bonds are always preserved and these cause the elastic properties. The dissolution of" all junctions would result in the death of the cytoplasm by liquefaction^ The great marvel of the Uving framework is its striking mobility,, which becomes apparent in protoplasmic flow. In this flow the chains, are orientated not only in small submicroscopic, but even in micro- scopic regions, as indicated by the visible strand formation. The paral- lel alignment of the chains is often so pronounced that birefringence of flow occurs (Ullrich, 1936a; amoeboid movement of the rhizo- podiae, Schmidt 1937a, 1941b). The whole movement is only intel- ligible if a great number of junctions are continuously being formed^, only to be broken down shortly afterwards. The jimdamental dijference from dead gels lies in the fact that in the cytoplasm the junctions are continuously reconstructed. The pattern of junctions in living matter is not rigid and fixed as, for instance, in gelatin or still more in cellulose gels ; its only permanent feature is its continual change ! The reconversion to the system of junctions proceeds according to- some definite plan about which we remain completely in the dark.. A temporary change in stability can also be produced artificially,, owing to the thixotropic properties of the cytoplasm (see p. 66). By mechanical means (pressure, shock) a reversible liquefaction can be brought about. Such drastic interference is always followed, however.,^ by a more or less serious damage to the cytoplasm (see p. 187). I CYTOPLASM 179 Interrelation of the particulate globules and the reticulate ground-cytoplasm. While it is fairly well established that the submicroscopic reticulate structure of the cytoplasm is formed by linear aggregation or by reversible denaturation of globular protein molecules, there is no proof that all existing submicroscopic protein particles participate in these sol- gel transformations. It is possible that certain globules, as, e.g., the microsomes in the liver, may be specialized for metabolic work, whereas others with the capacity of forming gels have the character of structural proteins. It seems unlikely that the two funda- mental tasks of the cytoplasm, metabolism and morphogenesis, are performed by the same globular elements. It is true that some investi- gators think of a uniform type of cytoplasm; thus Virtanen (1948) finds that the number of enzymes in bacteria is so high, that all protein molecules in the cytoplasm must be enzymes. On the other hand, we find that in the microscopic domain individuaUzed and mobile meta- bolic centres, such as erythrocytes or chloroplasts, are suspended in a liquid which can gelate (fibrinogen- fibrin transformation, sol-gel transformation of the endoplasm). Similar specialization might there- fore conceivably prevail in the submicroscopic domain. We may note here that pieces of cytoplasm separated from the rest continue to live independently, although they are not capable of restoring the original cell shape. Thus, since metabolism is confined to quite specific molecular configurations, all essential groupings have to occur repeatedly in each cytoplast ; this is the case if they are carried by submicroscopic particles. The development of the organism is presumably also governed by special specific groupings in the cytoplasm, which can be designated as morphogenetic configurations. However, in contradistinction to the majority of active groups regulating the metabolic process, these configurations do not by any means occur in every type of cell ; they are confined to the cells of certain tissues, probably located in the nuclei. A tissue of this kind acts as "organizer" (Spemann, 1936; Weiss, 1939; Baltzer, 1942), since the processes of development concerned can only take place in its presence. This organizer can be influenced by chemical means. Lehmann (1937a, b), for instance, has succeeded in controlling chorda formation by treating the gastrula of Triton or Rana with lithium chloride. This can be explained by as- suming that the essential morphogenetic configuration is changed l8o FINE-STRUCTURE OF PROTOPLASM II either substantially by chemical compounds (e.g., hydration) or only in its configuration in space (e.g., by changes in the distance between decisive groups) in such a way that they can no longer fulfil their task. These morphogenetic groups often require hormones to be activated (Hadorn, 1939). Since the morphogenetic faculties are assigned to special cells, whereas certain metabolic phenomena, such as respiration, are com- mon to all cytoplasts, a morphological separation of these manifes- tations of life in the submicroscopic domain is probable. As previously pointed out, the submicroscopic microsomes must contain a considerable number of protein macromolecules and other compounds such as nucleic acids, phosphatides, lipids, pigments, etc. These constituents must be united in some very specific pattern. This follows from the fact that their arrangement is capable of specific achievements in biosynthesis. Just as in organic chemistry an asym- metric synthesis is only possible if another optically active compound with asymmetric carbon atoms is present which prevents the form- ation of racemic mixtures, so, too, the organization of biocatalysts must be adequate to the chemical structure of the specific compounds synthesized. For here, as in the case of asymmetric synthesis, the theorem applies: Specific structures can he formed only hj the agency of corresponding structures. The chemical compounds of the cytoplasm would not be capable of accomplishing any useful work without definite positions in space. The prosthetic group (coenzyme) of an enzyme is only active when attached to a special protein carrier (apoenzyme). Although the chemical forces of their linkage are not considerable, and the coenzyme can therefore be split off and recombined with the macromolecular carrier with comparative ease, the system is only effective when the prosthetic group takes up its specific steric position. When the enzymes are located in individual particles such as micro- somes or mitochondria, they can be separated from the other cell constituents and examined in the isolated state. In the case of the endoenzymes, however, which cannot be extracted from the tissues (Bersin, 1939), the apoenzyme may be a part of the cytoplasmic framework, in which case there is, of course, no possibility of dis- tinguishing metabolic from structural cytoplasmic constituents. In connection with the foregoing it is necessary to stress the fact 1 CYTOPLASM l8l that morphogenetic manifestations of the cytoplasm are only possible in its gelated state, for this alone permits it to assume shapes different from those induced by the surface laws of liquids. Submicroscopic morphology is therefore very much concerned to know the type of junctions by which the macromolecules of the cytoplasm lose their individuality and aggregate to form a gel. Comparison with current opinions on the structure of cytoplasm. The views on the submicroscopic structure of cytoplasm developed in former editions of this monograph have met with some criticism. Before going into this criticism, we shall briefly discuss various points which make our theory fundamentally different from others. It is not permissible to draw a parallel between "protoplasmic vis- cosity" and the viscosity of liquids (compare Table XXII, p. 169). For here it is not merely a matter of friction between freely moving particles, but of an additional resistance offered by an elastic, sub- microscopic framework as well. I completely agree with Scarth (1927) when he writes that the fall of a particle through the cytoplasm is comparable to the zig-zag path of shot falling through a brush heap, and that drastic methods like centrifugation forcibly destroy the fine framework of the plasma structure. The work of Scarth also contains the essential points of this monograph in those places where he points out that the polarity and the capacity for growth of cells are incom- patible with the nature of a liquid such as that which has often been attributed to the cytoplasm and the nucleus. Often microscopic strands are visible in the cytoplasm. As a dense, tough, "formed" protoplasm, these are embedded in "unformed" protoplasm of semi-liquid consistency. Such differentiations have been distinguished as kinoplasm and matrix (Scarth, 1927), active plasma and paraplasm (v. Mollendorff, 1937) or spongioplasm and en- chylema (Monne, 1942a). In some cases the two constituents can be separated in the centrifuge as a gel rich in lipids and a sol, poor in lipids but rich in mitochondria, comparable to the conditions in the nucleus, where the chromosomal threads and the karyolymph can be separated from each other. The microscopic cytoskeleton (Peters, 1937) is not to be identified with the submicroscopic structure. Un- doubtedly the strands which are visible in the ordinary microscope originate from far-reaching bundling of the submicroscopic stra nds postulated by us, but they certainly are not homogeneous and poss ess l82 FINE-STRUCTURE OF PROTOPLASM II an invisible fine-structure, detailing of which falls within the province of submicroscopic morphology. A further task is to establish the nature of the plasma liquor (enchylema, paraplasm, matrix). Very many of the hypotheses relating to the structure of cytoplasm, discussed in former times (Lundegardh, 1922, p. 242), are irre- concilable with our own views. Nowadays the emulsion and alveolar theories can no longer be regarded as valid. Taking clotted milk as an example, Seifriz (1936) shows how the droplet theory takes ac- count only of the relatively coarse units, whereas the fine-structure is caused by the fibre structure of the casein. He applies this model to cytoplasm and is thus led to a scheme of protoplasmic structure which tallies well with ours, so long as we bear in mind that, when living, it does not represent a fixed coagulum of protein particles, because the particles may be reversibly released and move freely and independently of each other. Further comparison of the proteins of protoplasm with a heap of rodlets seems less felicitous to me, since such a heap has a fortuitous, statistical character, whereas the structure of protoplasm must be a co-ordinated whole. Its framework cannot be a disorderly pile; it must surely consist of an organized and well-defined structure. According to our present knowledge, all hypotheses of proto- plasmic structure which postulate permanently individualized sub- microscopic particles (granules, droplets, alveoles, ultramicrons) must be discarded as being corpuscular theories. The framework structure of gelated cytoplasm possesses no dispersed phase in the sense of the classical theory of colloids: both the framework and the enchylema are continous throughout the whole space available. For the same reason Butschli's foam structure or honeycomb theory cannot be taken into account, in spite of its numerous merits, for a honeycomb con- sists of closed dispersed regions in contradistinction to the open and continuous system of interconnected strands. Flemming's fibrillar theory, on the contrary, conforms rather well with the condition of a complete intermeshing of strands and dis- persing medium shown to be likely in this monograph. Here again, however, the fibrillar structure has to be transferred to submicro- scopic regions. In fact, in a three-dimensional network, both the contours of the meshes and the meshes themselves fill all space continuously. Monne (1946 a) is of the opinion that the protoplasmic fibrils do not form a network, but are only plaited (in German: Flecht- 1 CYTOPLASM 183 werk). To my mind this depends on whether we have to do with a plasma gel or a plasma sol (p. 1 65). In the first case there must be some interaction between the invisible fibrils, whereas in the second case they may be independent of each other. The fibrillar theory has been developed partly on the basis of fixed ■structures. This derivation is not as unreasonable as has often been suggested, since on fixation the submicroscopic or amicroscopic strands of the cytoplasm combine into coarser strings by directed ^coagulation and can thus become microscopically visible. It is only because the cytoplasm actually possesses a thread structure, that the good fixations obtained by cytological micro-techniques are possible. In this process the molecular framework may shrink, be coarsened, ft * h n 0 ^-„„-d ^) ^) Fig. 123. Microscopic chromosome structure (from Heitz, 1935). a) Idealized chromo- some with helicoid chromonema threads; heterochromatic region hatched; in the upper part : a primary (kinetic) constriction ; in the lower part at the right : secondary constriction with satellite (corrected to satisfy Geitler's criticism 1938, p. 98). b) Chromatid pair of Trillium erectum. c) Spiral structure of the chromosomes of Tradescaiitia virginica. d) Spiral structure of the chromosomes of Trillium erectum. the structure of chromosomes can be built up from the chromonema theory. Each chromosome contains two, according to other investi- gators (Nebel and Ruttle, 1937; Nebel, 1939, 1941) even four spirally wound threads, called chromonemala (Fig. 123 a). In the first case they are identical with the chromatid threads (Fig. 123 b), well known from the prophase of meiosis. It is only in that prophase that the chromonema spiral is completely uncoiled and therefore survey- able in its entire length, which is many times that of the chromosome. It consists of a non-staining thread which at regular intervals is covered with knots showing the nucleal reaction and designated as chromo- meres. In the mitosis chromosomes these particulars can scarcely be observed, as the chromonemata are coiled (Fig. 123 c) and embedded 220 FINE-STRUCTURE OF PROTOPLASM II in a ground mass (matrix) which shows strong nucleal staining. It must be mentioned that a spiral structure has been observed onl)^ in large chromosomes. According to Japanese and American cyto- logists (Straub, 1938; HusKiNS, 1 94 1, 1942) the visible helix some- times possesses a spiral structure of its own, in which case the chro- mosome would possess the structure of a doubly wound helix with a primary and a secondary spiral (large and small spiralling). With the aid of the phase contrast microscope Ruch (1945) has shown that in the case of the much-investigated chromosomes of Tradescantia the chromomeres occurring in pairs on the spirally wound chromonema fibrils suggest the existence of the small spiralling; but by judicious considerations of the focal depth of the microscope objectives used, he proves clearly that no doubly wound chromosomes exist (Ruch, 1949). The question as to how the helical chromonemata are separated from each other during mitosis without uncoiling is a problem in itself (Matthey, 1941). The chromonema theory has gained general importance by the discovery of the giant double chromosomes of the nuclei from the salivary glands of the Diptera. In these marvellous cytological objects homologous chromosomes are united into astonishingly broad and remarkably long ribbons. These gigantic chromosomes may be re- garded as bundles of numerous expanded chromonemata, formed by endomitosis (Heitz, 1935). They are united into strings of consider- able dimensions ; the chromomeres form transversal discs which, by means of staining or the nucleal method, are made visible as numerous crosshnes (Fig. 124). The non-stainable, anucleal regions of the chromonemata bundles represent the protein components of the chromosome. It may be concluded that the protein thread is not restricted to the colourless segments, but runs invisibly through the whole chromosome. In the chromomeres the nucleic acid components are localized, thus masking the protein ground mass. Their localization is demonstrated by the nucleal reaction, the ultraviolet absorption and the X-ray absorption method of Engstrom which proves that the Feulgen positive bands contain 2 to 10 times more mass than the Feulgen negative ones.^ Moreover, the ultraviolet microscope with its higher resolving power furnishes proof of the existence of the protein ground mass in the ^ Engstrom and Ruch (1951). NUCLEUS 227 chromomeres. With the aid of digestion experiments in which the nucleic acid was protected from digestion by lanthanum thymo- nucleate, Caspersson (1936) finds that the chromomeres are resolved into extremely thin discs. Ultraviolet photography reveals a fine- structure of lamellae with a thickness of only 0.1 j.1. Since at this order of magnitude the limit of the resolving power in ultraviolet hght is reached, the question as to whether these very thin chromomere discs possess a still finer submicroscopical structure and thus are subdivided remains unsettled. Personally I do not doubt that they are. Conversely, milt nuclease digests the nucleic acids of the chromomeres (Mazia and Jaeger, 1939) without disturbing the ground structure of the chromosomes of the salivary glands. The ability to take the Feulgen stain disappears; on the other hand the ninhydrin test turns out positive over the entire length of the chromosome. So the chromonema does not consist of alternating protein and nucleic acid links, but represents a continuous protein thread in which at regular intervals nucleic acid knots are inter- calated. The nucleic acids form saltlike compounds with the protein ground mass, the nucleoproteins, whose oc- currence is therefore limited to the chromomeres (Fig. 125 b-d). Fibrillar hypothesis. From a morphological point of view Wrinch (1936) believes the molecular structure of the chromonema to be as follows : the polypeptide chains form a system of parallel fibrils like the warp of a weaving-loom and the nucleic acids represents the woof in this system of chains. Every four neighbouring polypeptide chains are kept together by a molecule of the tetra-basic thymonucleic acid. Fig. 124. Two incompletely conjug- ated giant chromosomes of the nuclei from the salivary gland of a Drosophila hybrid with a chromosome pattern characteristic of the two parental species (from Patau, 1935); mel from Dr. melaiiogasier, sim from Dr. simu- lans ; in a a. structural difference. 228 FINE-STRUCTURE OF PROTOPLASM II As the native nucleic acids have a much higher molecular weight, the woof would not consist of short chains with four members but of much longer chains. The heteropolar salt bonds between the acid groups of the nucleic acid chains and the basic groups of the poly- peptide chains would have to be considered as the junctions of this network (Fig. 125 b). »t = =2-t Biti?-t4 h) iMm d «') Fig. 125. Submicroscopic arrangement of nucleic acid (shaded); a) in the head of the spermatozoon of Sepia (after Schmidt, 1937a); b-d in the chromonema: b) transversal (after Wrinch, 1936), c) lengthwise (after Schmidt, 1937c, 1939b), d) scattered orienta- tions (from Frey-Wyssling, 1943b, 1944a). This scheme is not supported by the optical properties. The arti- ficial nucleic acid threads obtained by Schmidt (1957a) are optically negative^ ; and since in the spinning the molecular chains are arranged parallel to the axis, the polynucleic acid chains themselves must also be optically negative. It follows from this that the molecules of the nucleic acid chains in the sperm nuclei (Schmidt, see Fig. 125 a) run parallel to the morphological axis of the sperm head. But the poly- peptide frame of these nuclei also must be orientated in the same direction. This means that the chain molecules of both nuclear com- ponents show parallel alignment. The chromomere discs of the chromosomes of the saUvary glands are optically negative (Schmidt, 1937c, 1959b). For the submicro- 1 Threads of sodium thymonucleate show a reversal of their sign of double refraction when strongly stretched (Wilkins, Gosling and Seeds, 195 i). Z NUCLEUS 229 scopic fine-structure of the chromomeres Schmidt therefore takes into consideration a possible arrangement as given in Fig. 125 c. The fact, ascertained by Astbury and Bell (1958), that the fibre period of 3.34 A of the nucleic acids is about the same as that in the poly- peptide chains (3.5 A, see Table XXXII, p. 368) seems to support this hypothesis. With the aid of the ultraviolet dichroism of the nucleic acid chains Caspersson (1940b) has checked the structure proposed in Fig. 125 c. If the nucleic acid molecules in the protein fibres showed complete orientation, chromomeres, like artificial thymonucleic acid fibres, would displav a very pronounced dichroism in polarized ultraviolet light. Compared with these fibres, however, the chromomeres of the chromosomes of the salivary glands show only an extremely small dichroitic effect. Caspersson therefore draws the conclusion that the nucleic acid chains are intercalated practically without orientation. Also the double refraction of the chromomeres, as derived from the birefringence of flow of sodium nucleate sols (Signer, Caspersson and Hammarsten, 1938), proves to be very small. Meanwhile, as- suming that nucleic acids are straight chains, the negative sign of the chromomere birefringence indicates that the chains have a certain preferred orientation. With the aid of the formula mentioned on p. 90 the scattering in the orientation of the chain molecules can be calcu- lated (Frey-Wyssling, 1943b), and the scattering angle found in this way is 86°. 5, i.e., nearly a right angle. This means that the scattering is almost complete, thus furnishing an important argument against the supposition that the nucleic acid molecules are parallel to the chromonema axis. A similar result is obtained if the intrinsic double refraction of — 0.050 found by Schmidt (1928) for the chromatin of the Sepia sperm, or even only a fraction of this value, is compared with the birefringence of the chromomeres. In spite of the small orientation of the nucleic acids, Caspersson assumes the protein chain structure to be continuous. Orientated polypeptide chains are supposed to cause the anucleal chromosome segments to appear positively birefringent. This effect, however, can only be observed in stretched chromosomes (F. O. Schmitt, 1938; Pfeiffer, 1 941 a). Optics therefore do not provide sufficiently reliable data to assume orientation of the protein chains. There can certainly be no pronounced fibrillar texture of expanded polypeptide chains. 230 FINE-STRUCTURE OF PROTOPLASM II as there is not the slightest indication of the existence of a chain lattice. All the same, the anisotropy of swelling and the cleavabiHty of the chromosomes are in favour of an orientation along the long axis of the protein ground mass. In his microchemical experiments with chromosomes of salivary glands Painter (1941) is also impressed by their fibrillar character. It must be supposed that the isotropy of the protein results from coiling and folding of polypeptide chains, com- bined with a corresponding hydration. Evidently the chromosome protein consists rather of globular molecules which may be aggregated to form beaded chains. The explanation of the birefringence of other chromosomes (Schmidt, 1957a, 1941a) is much impeded by our insufficient knowledge of the submicroscopic orientation in the chromonemata bundles of the salivary glands. According to Becker and Kozbial (1937), the optical character of the chromosomes of the root tips of Allium and Vicia depends on the process of fixing : if treated directly with alcohol they appear to be negative; after a previous treatment with acetic acid vapour (causing swelling) they are positive. On the assumption of a nearly complete scattering of the nucleic acid chains in the isotropic living chromosome, these effects might be explained by tendencies towards orientation as a result of the shrinkage or swelling in the fixing process. It seems to me that considerations of this kind open more prospects than explanations formerly attempted with the aid of the spiral structure (Nakamura, 1937). Kuwada and Nakamura (1934) explain the positive double refraction of the chro- mosomes of Tradescantia by a single spiral of negative chromonemata ; whereas, in their opinion, optically negative chromosomes are caused by a double- wound spiral. d. Submicroscopic Morpholo^ of Hereditary Processes Genes. The fibrillar character of the chromatids meets two important morphological requirements of genetics: i. the substrate is easilj ckavable in the direction of the long axis, which is not only necessary for the splitting of chromosomes but also for the phenomena taking place between synapsis and diakinesis in heterotypic divisions; 2. the long chromonemata offer an opportunity for the linear arrangement and the possibility of exchange of the genes. AIorgan's school has calculated that the number of genes known 2 NUCLEUS 231 in the Drosophila chromosomes is so large that for reasons of space each gene must be bound to relatively small molecules of about the same order of magnitude as found in the reserve proteins (compare Fig. 90, p. 136) investigated in Svedberg's ultracentrifuge. It is difficult to see, however, how such freely moving particles are able to intervene decisively in the processes of development. To be able to do this their carriers must have fixed mutual positions and it is best to imagine that they are fixed on beaded protein fibrils. In this way we comply with the requirement of Hnear arrangement in a manner which could hardly be improved. In spite of its great probability, however, irrefutable proof of the existence of the submicroscopic fibrillar structure has not yet been produced. As has been shown, the quantitative evaluation of the optical results suggests that the protein is not in a pronounced fibrillar state possessing the characteristics of a chain lattice ; and thus far the electron microscope has failed, because even the pachytene and diplo- tene chromosomes yield only compact black shadows (Elvers, 1943) showing fewer particulars than a good light optical image. It is of so much the greater value that the experimental investigation of mutation or ray genetics (Zimmer and Timofeeff-Ressovsky, 1942) opens new perspectives. Target theory. Artificial mutations are induced by ionizing rays (UV- rays, X-rays, y-rays). The dose of rays is measured by the X-ray unit r (roentgen), which is defined as that amount of rays which will bring about enough conductivity under prescribed conditions in a chamber of I cm^ of air to permit a charge of one electrostatic unit to be measured at saturation current. It is now established that the mutation rate induced artificially by radiation is proportional to the dose of rays brought to bear (Timofeeff-Ressovsky, 1940). The effect is independent of the wavelength and the dose (intensity X time) can be given all at once, concentrated or diluted, or else given at intervals. There appears, therefore, to be no recovery. For instance, whereas the sex-linked mutations in the x-chromosome of Drosophila have a natural mutation rate of approximately 0.2%, the irradiation of 2500 r' produces a rate of 7% and 5000 r produces 14^%. If the mutation rates are plotted as a function of the dose, a straight line resuhs which intersects the zero point; thus there is no threshold value and any small dose will give an effect. 232 FINE-STRUCTURE OF PROTOPLASM II Ionization consists in the formation of ions from neutral molecules by the action of irradiated energy. The molecules in question are, as it were, struck by the energy quanta of the radiation and are thereby modified. That is why the occurrence of a single ionization is called a hit. The relation between mutation rate and dose of rays indicates that a mutation is the result of such a hit. It can also be demonstrated (Timofeeff-Ressovsky, 1940) that the interdependence of dose and rate would not produce a straight line if several hits were needed to bring about one mutation. The conclusion to be drawn from the bio- physical analysis of chromosome irradiation is, therefore, that the artificial mutation of genes is the elementary result of a single hit. There are, it is true, other possible physical explanations, besides the target theory, which may account for the effects observed (Minder and Liechti, 1945). The approximate number of atoms in a cubic centimetre of organic sub- stance being known, and also the number of single ionizations which one r unit is able to evoke, it is possible to calculate how many atoms are needed for one of them to be hit to produce the mutation in question, this by means of the experimentally ascertained mutation constant, which indicates the degree of probability to incite a mutation by a given dose of radiation. The volume occupied by these atoms altogether is called the target area. It varies with different mutations of genes within the x-chromosome of Drosophila; nevertheless an average can be calculated, according to which the susceptible volume amounts to 3.20-10-20 cm^., from which it follows that the radius of the target area (assumed to be spherical) is 1.97 m/x (Timofeeff- Ressovsky, 1940). There is an alternative method by which the target area can be computed. If very strong ionizing rays are used, of very great density, such as neutron rays, for example, more than one ionization may take place in one target area, only one of which, however, effects mutation. The other ionizations are inoperative and the mutation rate must consequently be smaller than was to be expected from the irradiated dose of rays in r units. Indeed, in the case of the >r-chromosome of Drosophila, the mutation rate actually is 1.6 times smaller for neutron rays than for X-rays, with the same dose of rays. The radius of the spherical target area can now be deduced from this factor to- gether with the known density of ionization for neutron rays; Lea (1940) finds 1.89 m/i,. Seeing that this figure so nearly agrees with the value found by Timofeeff-Ressovsky, it may be taken as fairly certain that the order of magnitude of the target area is roughly 4 m/x diameter. The target area is not to be identified with the gene, since it only gives us the size of a sensitive area within which something has to happen favouring the probability of a mutation. The gene may therefore be larger than the 2 NUCLEUS 255 target area, viz., if not all parts of the former are capable of changing their molecular structure by ionization. It may, alternatively, be smaller, if it does not attain the overall dimension of the statically calculated ionization area which, according to Timofeeff-Ressovsky (1940), contains 100-2000 (with a mean of roughly 1000) atoms. The latter possibiHty is, however, discounted by estimations of the size of the gene made by speciaUsts ingenetics.True.it is often stated in the literature that the target area is of the same order of magnitude as that of the gene size found by other methods, but we shall show that this is not so. One known estimation of the kind comes from Muller(i93 5). Assuming that a single chromonema thread of the salivary gland chromosomes had the same volume as in the corresponding metaphase chromosomes of normal cells, the following calculation applies to the x'-chromosome of Drosophila. In metaphase its volume is 1/8 /x^, two-thirds of which fall to the share of the chromonema, the length of which in the salivary gland chromosome is 200 ^. When completely uncoiled, therefore, a single chromonema thread has the submicroscopic thickness (cf. Metz, 1941) of 0.02 /x. The thread is thinner still if it is assumed that the chromonema is regularly screwed-up in the metaphase chromosome, the diameter of which is \ [x. The length of 200 fi gives us 250 windings; consequently, with the chromosome being 2 fx. in length, the chromonema could not be thicker than 0.004 /x. In calculating the length of the chromonema section containing a gene, MuLLER was guided by the following consideration: By examining the interchange of factors in cross-breeds, four genes were localized in a given chromomere of 0.5 ju. width in the salivary gland chromosome and the existence of further genes was shown to be improbable. Thus the length covered by a gene on the chromonema thread would be about o.i 25 ^. This is a dimension which lies on the borderline of microscopic resolving power. The chromonema sections which, according to Muller, correspond ap- proximately to one gene, are shown in diagram in Fig. 1 26d and, for com- parison, the target area is indicated by a black circle. It is recognized that the thickness of the chromonema thread is of the same order of magnitude as the diameter of the target area, but never the estimated size of the gene, the volume of which exceeds that of the target area by two to three orders of magnitude ! It can be shown that the sphere of action within a gene has a similar size to the target area. Carrier hypothesis (Frey-Wyssling, 1944b). If the volume of the gene is liable to be more than a thousand times larger than the target area, what, it must be asked, are the relations between these two quantities? It will be seen in Fig. i26d how the small, sensitive region is embedded in a large, non-mutating area. It is not known where the sensitive region lies and it may therefore, if desired, be thought of as placed anywhere. The picture is reminiscent of that of enzymes ^ 234 FINE-STRUCTURE OF PROTOPLASM II where small active groups are similarly carried by a larger protein complex system (see Fig. 119, p. 208). I therefore propose to discuss the picture of carrier and prosthetic region for the genes as well and to call this view the ^''Carrier hypothesis^'' . Since it is not the salivary gland chromonemata stretched to the utmost, with their hypothetical submicroscopic thickness, which operate in hereditary processes, but the considerably shorter meiotic chromosome threads (leptonema, zygonema), the size of the gene should be derived from the conditions produced by reduction division. We learn from genetics that the regions inciting mutation are placed linearly in the conjugating chromatids; consequently the target areas must likewise be aligned lengthwise in the leptotenic chromosomes. As the x-chromosome is supposed to contain about 1800 genes (Timofeeff-Ressovsky, 1940), all sensitive regions of 4 m^^ diameter should together produce a length of 7.2/^. Bearing in mind that, according to Muller (1935), the genetically active volume of the x-chromosome is roughly ^12 Z*^? one finds for the thickness x of the extended leptotenic chromosome prior to conjugation of the chro- mosomes, ^l^-K^-Tz-j.z = 7i2) from which x = 0.12 /^ is derived for the thickness of the so-called leptonema. This value may be of the right order of magnitude, since the diameter of the leptonema is in the vicinity of microscopic resolving power. If the genetically active chromomeres of the leptonema are now divided up into submicroscopic slices of the thickness of a target area, the region corresponding to a gene should be included. In this view and by this calculation, the gene should be a flat disc having an estimated diameter of 120 mju and thickness of 4 m^. We are com- pletely ignorant as to where the target areas lie in these discs: the arrangement may be any of an almost unlimited variety, as in Fig. 1 26 c. The only certainty we have is that, given the linear ahgnment of the loci of the mutations, they must be juxtaposed in the axis of the leptonema. In Fig. i26e such a gene disc is represented on the same scale as the dimensions of the gene calculated by Muller (1935). The position of the target area within the gene being unknown, it is shown as a globule placed arbitrarily anywhere in the disc. It is interesting to note that this size of the gene tallies well with that computed by Muller (chromonema cross-section x lengthof gene), although found by total- ly different means (leptonema cross-section x diameter of target area) : NUCLEUS 235 Gene size according to Muller(i93 5), Fig. izGd (2000)- 50,000 (m/^)^ Gene size according to carrier hypothesis. Fig. 126c 45>ooo (m^)^ Target area according to Timofeeff-Ressovsky (1940) 32 (m^)^ Chromosome Gene 1.7m u lOmu r Fig. 126. Chromosome and genes, a) -v-Chromosome of the Drosophila; b) leptonema with linearly aligned chromomeres; c) leptonema, strongly magnified, with interchromomeres and target areas indicated as points; d) size of genes calculated as prisms with quadratic cross-section according to Muller (1935), each containing a target area in corresponding size, di upper and dg lower limit of gene size; e) gene size according to the carrier hypo- thesis, at the same magnification as d; e^ front view of the gene disc in comparison with a target area, e, ground-plan of the gene disc, indicating the twofold construction out of 2 chromatids with equivalent spheres of action;/) yellow respiratory enzyme with apo- and co-enzvme, both drawn on the same scale as d and e. Scheme i26e is even more reminiscent than i26d of the structure assigned to enzymes. Fig. i26f depicts the yellow respiratory enzymes on the same scale. Both the size of the colloidal carrier (mol. wt. 70000) of this enzyme and its prosthetic group are known. Presuming that 1000 atoms occupy a volume of 3.2- io"20cm^ (Timofeeff-Ressovsky, 1 940) and that, in accordance with the composition of sturine (Fig. 121, p. 212, 27 C : II N : 5 O : 47 H), the average weight of the atoms of the amino acids is 6.7, the diameter of the apo-enzyme (thought of as a sphere) is calculated to be about 10 m/^ and that of the co- enzyme with 81 atoms, approximately 1.7 m/*. 236 FINE-STRUCTURE OF PROTOPLASM II It will be clear from the following that a similar comparison applies to other enzymes, at any rate so far as the prosthetic group is con- cerned : Co-enzyme of carboxylase (aneurinopyrophosphoric acid) 44 atoms Co-enzyme of the dehydrogenase II (nucleotide of nico- tinic acid + phosphoric acid + nucleotide of adenine) hydrogenated -^ 7^ ,, Co-enzyme of the yellow respiratory enzyme (lactoflavin- dinucleotide of adenosine) 81 ,, Average target area of the genes 1000 ,, The apo-enzymes of these desmoenzymes are not freely moving colloidal particles; Hke the genes in the chromosome, they are em- bedded in the submicroscopic cytoplasmic structure. Only by auto- lysis can they be liberated under certain circumstances and made accessible to examination. The comparison between gene and enzyme may not be merely a superficial one ; one might at least try to probe further. It is scarcely to be wondered at that the gene and target area should be so much larger than the volume of the apo- and co-enzymes, if it be remem- bered how much more complicated than single metabolic reactions are the processes of development controlled by the genes. Latterly it has become ever more evident that this control is exercised chemically. When a mutation takes place, these chemical processes proceed differently. It is therefore not wrong to assume that the target area acts like the prosthetic group of an enzyme and that the controlled processes follow a different course owing to changes in this sensitive area. As this area contains approximately 1000 atoms, 50 amino acid residues (with on an average 20 atoms) are located in it, allowing protein chemistry to come into full play in its almost unlimited variety. According to the hypothesis propounded here, the gene would, in the terminology of Haase-Bessel(i936), consist of a carrier (pheron) and chemically active regions (agon), some idea of the dimensions of which can be formed on the basis of the target theory. Since, however, every colloidal particle of the apo-enzyme carries only one amicro- scopic operative group, this conformity probably cannot be assigned to the genes. There are, maybe, several chemically active regions in the large disc of Fig. i26e. This would explain polyphaeny, i.e., that 2 NUCLEUS 237 phenomenon whereby often more than a single property is regulated from one locus of the chromosome. This view would also account for minor distinctions in activity of homologous active regions in different individuals and would explain polyallely. There must be some correlation between the carrier and the active region, in the same way as a co-enzyme can only develop its activity in close conjunction with the apo-enzyme; to some such reciprocal action must be attributed the different ways in which certain phaena are actualized. To be brief, the carrier model serves to make intelligible most of the knowledge acquired by research into heredity. There being good grounds for assuming that the leptonema has a double structure, falling into two chromatids after conjugation, the carrier discs may be represented as two halves, each with an operative region of the same value (Fig. iz6e). Of these two, only one need be struck by the rays for the origin of a mutated gamete, since the chromatids are separated from each other in the formation of tetra- cytes. The advantage which the carrier hypothesis possesses, as compared to th.Q fibrillar hypothesis developed in the chapter on chromosomes, is that it disregards the disputed question of molecular protein structure and nucleic acid intercalation ; the elementary units may be conceived of either as fibrillar protein units coiled in any way, or as globular proteins. Then, the carrier hypotheses makes the gene and operative region of suhmicroscopic dimensions, whereas the fibrillar hypothesis allows the gene to be of amicroscopic size represented by side chain groups of polypeptide chains. In the first edition this picture un- warrantably simplified the exceedingly complicated facts. On the other hand, the fibrillar hypothesis has to its credit the plausibility it confers upon the shape, cleavability and self-duplication of the chro- mosomes. It will therefore be the aim of research to reconcile these two hypotheses to a concordant theory by endeavouring to fathom the suhmicroscopic morphology of those proteins which represent neither their extreme fibrillar, nor an independently dispersed globular form. Function of the desoxyribose nucleic acids. The desoxyribose nucleic acids, which were at first thought to be the hereditary substance par excellence, are of relatively uniform chemical constitution and, in their molecular morphology, lack the diversity required by genetics. 238 FINE-STRUCTURE OF PROTOPLASM II Moreover, Caspersson's measurements show that their appearance is transitory and that they afterwards largely disappear. For this reason Kiesel (1930, p. 185) stigmatizes as downright paradoxical the fact that cytologists pay such conscientious attention to an unspecific material like the desoxyribose nucleic acids, yet ignore the proteins, with their specific structure, merely because constituents do not bind the basic dyes used for staining cell nuclei. Posternak (1929) goes to the length of relegating the nucleic acids to the rank of degradation products of organic phosphorus compounds ; but this view is invalid- ated by the morphological behaviour of the desoxyribose nucleic acids during karyokinesis and the interesting fact that many co- enzymes consist of nucleotides (co-dehydrogenase II and others, see p. 208). I have therefore suggested the following hypothesis respecting the function of the desoxyribose nucleic acids : The genes play no active part during karyokinesis, but 2it& passive and in this state are distributed by some process among the daughter nuclei. Their operative groups must therefore be reactive in the active nucleus to fulfil their task, but they must be screened off during nuclear division. This might be effected by a loose binding of desoxyribose nucleic acid groups. It was pointed out in the discussion of the phosphatides that, in the respira- tory combustion of carbohydrates, those hydroxyl groups of the sugar which are not subject to degradation are screened by phosphorylation and are thus temporarily protected. Similarly, the phosphoric groups of the nucleic acids might for the time screen the specific groups of the genes during mitosis. This would account for the localization of the desoxyribose nucleic acids in certain places only, viz., where the genetically active groups are to be found in the fundamental protein substance. They thus give a true picture of the distribution of genes as proved by cytology. There is, then, nothing "paradoxical" about the attempts to establish the distribution of the desoxyribose nucleic acids in the chromonema down to the finest detail, since these are the indicators, as it were, of the specific groups through which the genes operate. The assumption that the desoxyribose nucleic acids accumulate only in those parts which contain genes, and protect their active groups, integrates the conflicting views championed by the theorists of he- redity, one being founded on the structural chemical specificity of the 2 NUCLEUS 259 proteins, whereas the other side upholds the micro-morphological specificity of desoxyribose nucleic acid distribution. The fate of the desoxyribose nucleic acids in the cycle of nuclear division favours the above hypothesis. When the nucleus undergoes mitosis, desoxyribose nucleic acids are built up (increasing chromo- philic tendency, nucleal reaction and ultraviolet absorption). In pro- phase they appear to be embedded in the chromomeres, protecting the specific groups during the cleavage of the chromosomes. When their task is done, most of the desoxyribose nucleic acids migrate from the chromomeres to the matrix of the chromosomes. As a result, the latter absorbs stain to the full extent and the chromonemata thus remain invisible during metaphase and anaphase; in this stage, there- fore, nothing at all can be known of their exact morphology. In telophase the desoxyribose nucleic acids are for the most part degraded again. The chromosomes become transparent and it can be seen how the chromonemata, losing their stainability, uncoil (Heitz, 1935, p. 419) and disappear in the nuclear reticulum. There are parts of certain chromosomes which are called hetero- chromatic, where the desoxyribose nucleic acids are not degraded after cell division. When genes of euchromatic regions of the chromosome come into the neighbourhood of heterochromatin by crossing over, their manifestation is lost (Prokofyewa Belgorskaja, 1948; Lewis, 1950) or changed from dominant to recessive (McClintoc, 1950). These facts indicate a screening effect of desoxyribose nucleic acids on genes. In the view set forth here, the desoxyribose nucleic acids play a passive part in heredity, in that, although they protect the genes, they do not participate in their spontaneous propagation. By contrast, on the analogy of the enzymes with nucleotides as prosthetic groups, an active part may be assigned to them. Caspersson (1941), applying his ultraviolet absorption method, discovered that vigorous protein synthesis is initiated wherever nucleic acids appear; notably that histones are formed as the result of the reaction of nucleic acids of the ribose type (absorption maximum at 2900 A) and globuhns from that of the nucleic acids of the desoxyribose type (absorption maxi- mum at 2800 A). Caspersson, therefore, declares nucleic acids to be necessary to any and every biological synthesis of proteins. In this case the desoxyribose nucleic acid would be operative in the redupli- 240 FINE-STRUCTURE OF PROTOPLASM II cation of the chromonema threads during cell division; but then the question arises as to why protein synthesis is only necessary in the chromomeres and how the anucleal parts of the thread augment their protein substance. According to this theory, nucleic acids would be also temporarily necessary in endomitotic division (Geitler, 1940; Berger, 1 941), though hitherto this has evaded observation. In what- ever way the function of the nucleic acids as synthesizing protein enzymes may be confirmed or modified in the future, it will not irre- concilably contradict the propounded hypothesis of screening, as in both cases nucleic acids must be assumed to accumulate in the genet- ically active regions, as a result of which the chemical activity of the genes is, for the time of multiplication, paralyzed. ScHULTZ (1941) goes one step further and calls the genes nucleo- proteins, that is to say nucleic acid compounds. He declares that the genes and nucleoproteins have in common the properties of speci- ficity, auto-reproduction, similar distribution in the cell and intimate relation to synthesis processes. There is this much to be advanced against this opinion : that the activity of the genes only begins in the reconstituted nucleus, whereas in that state the nucleoproteins dis- appear very much into the background. Hence, after their duplication and division, the genes must be independent, to a large extent, of the nucleic acids, making their influence felt in the growing cell, without having the character of nucleoproteins. Identical auto-reproduction of nucleoproteins (comparison with virus protein). Whereas in this monograph the genes have been compared morphologically and chemically with enzymes, the literature inclines rather to draw the analogy with the rod-shaped virus particles, notwithstanding the fact that important points of comparison have lost cogency since the invalidation of the classical fibrillar hypothesis of the chromosome structure. Many of the varieties of virus isolated so far are of similar chemical composition to chromatin: they are «»r/i?o/)ro/f/«j-, i.e., proteins of polypeptides and nucleic acids. They do, it is true, still contain lipids and, under some circumstances, also small amounts of polysaccharides. Minute amounts of lipid have also been detected in chromosomes (Hirschler, 1942), though as a rule those components are disregarded in discussions on the structure of chromatin. It is the virus of tabacco mosaic disease which has been subjected to the most exact analysis, as Stanley's method (1958a) provides a suitable means (by precipitations) of obtaining it in a crystallized form. It contains 1.7 to 5 % of nucleic acid, according to its preceding treatment. If the nucleic acid is separated off, the virus protein loses its pathogenic properties and 2 NUCLEUS 241 its propagating power. This proves beyond doubt that the mysterious auto- reproduction of the crystalhzable viruses is determined by nucleoproteins. There is, however, a fundamental difference as compared w4th the nucleo- proteins of the nuclei of the cell, the virus protein showing no nucleal reac- tion. Thus the phosphoric compounds in the viruses are of the ribosenucleic acid type, and not the thymonucleic acid found in the nuclei. The tobacco mosaic virus molecules are threadlike, judging by their birefringence of flow (Takahashi and Rawlins, 1933, 1935) and as demonstrated by the electron microscope (Fig. 84c, p. 126). The thread molecules unite into bundles liable to grow to microscopic dimensions and then appear as crystallized virus protein. This, however, is not in a true crystalline, but rather in a mesomorphous state, for the X-ray analysis of these "crystals" produces only intramolecular interferences (Bernal, 1939) and does not reveal any molecular lattice arrangement of the virus molecules (Wyckoff and Corey, 1936). Thus, like liquid crystals, the parallelized thread molecules are free to revolve and shift individually. The structure of the mesomorphous virus rodlets, which is reminiscent of that of the chromonema, favours their cleavability. On the other hand, the reduplication of the chromomeres can hardly be understood as a mere split- ting of bundles of parallelized molecules. The comparison is also prejudiced by our complete ignorance as to how the nucleic acids are distributed in the submicroscopically visible virus molecule. The analogy rests merely upon the common filiform structure. It is the mysterious auto-reproduction of the virus protein which encourages comparison with the chromonemata in the chromosomes. If only a trace of the thread molecules of tobacco mosaic virus finds its way into the cells of the tobacco leaf, they fill up completely, in an astonishingly short time, with the pathogenic protein, which becomes visible as birefrin- gent rodlets, whereas the protein proper to the cell diminishes. Thus, when in contact with virus molecules, non-virus protein becomes virus. This phenomenon has been termed autocataljtic reproduction. It is known in other compounds; for example, small amounts of trypsin are liable to change a larger amount of another compound, known as "protrypsin", into trypsin. Energy is required for the spontaneous reproduction of the virus protein and this is supplied by the living cell. There can, therefore, be no reproduc- tion of virus outside the living cell. It is tempting to regard the duplication of the chromonemata in mitosis likewise as autocatalytic reproduction; but we should not forget that we have simply coined a term for what is at present an inexplicable process and are still quite in the dark as to the nature of the "first step" which, through contact with the specific nucleoproteins of the chromonema, has autocatalyt- ically to be transmuted into identical nucleoproteins. The electron microscope shows that the rod-like shape of the tobacco mosaic virus (Wyckoff, 1947a) is an exception. The majority of the virus species photographed by Wyckoff (1947b) have a pronounced globular 242 FINE-STRUCTURE OF PROTOPLASM It shape and agglomerate in a visible crystal lattice. The morphologicaE analogy of chromonemata and virus, therefore, is no longer supported; the- chemical comparison of both genes and viruses with enzymes is much more convincing. Nucleus and cytoplasm. Considered from the morphological stand- point, the secret of karyokinesis is evidently that the specific protein- molecules, which serve as substratum to the genes, have to be carefully transmitted to the daughter cells, preferably without any reciprocal! changes of position along the chromatid. Their individuality and specific spatial relationships were developed in the course of phylo- genesis and the cytoplasm is not capable of re-creating them. The great riddle of heredity therefore still is : How can a chromonema of such complicated submicroscopic and amicroscopic morphology that it can never be produced anew, bring forth its like from itself by longitudinal division? This mysterious process must undoubtedly take place frequently in the giant chromosomes of the Diptera, which are bundles of similar chromonemata. It is as though the chromonemata served^ as it were, as patterns for the creation of their like. It is known from the evidence of the asymmetrical C synthesis (see p. 207) that certain configurations are able to produce essentially the same morphological forms in the amicroscopic region, but the refinements of this process, and its mechanism are a mystery. For here, as contrasted with the mode of action of the enzymes, it is not merely a question of fitting a key to a lock, but of how the key produces one identical to it, or the lock its exact like. If we take the specific structures to be a given fact, we come to an important decision as to the morphological signification of the nucleus. The gene-bearing protein threads are in a sense self-contained and irretrievable structures and it therefore becomes clear why they are not carried along by the cytoplasmic stream, but are localized at a given spot. There they are withdrawn from the turbulent activity of the cell and perform their directive and formative task as static centres. It is evident from the heredity of cytoplasm (Wettstein, 1957) that specific groups must also occur in it. These special structures, how- ever, are not solitary, for parts of the cytoplasm are similar in their behaviour to the whole cytoplast. Even fragments of the eggs of sea- urchin without a nucleus can undergo a certain development involving 5 CHLOROPLASTS 243 cell division (E. B. Harvey, 1936). If, on the other hand, portions of chromosomes are removed from the nucleus while division is going on, the result is a serious modification of the hereditary process. Although the cytoplasm is able to build up very complicated molecular systems, its architectural capacities are to some extent limited, for it cannot produce from itself the protein structures of nuclei and plastids. In heterotrophic organisms it even lacks the capacity to manufacture relatively simple elementary units, which are needed for protoplasmic synthesis; it is for this reason that these compounds have to be added as vitawins to the culture medium (ScHOPFER, 1936/37). As a rule, all such problems are studied in their purely chemical aspect. Yet the molecules should not be considered only as chemical supporters of reactions, but also morphologically as elementary units of the high polymeric gel frame. In the cytoplasm, this texture is very finely spun, is labile and is involved in permanent reconstruction. In the chromosomes of the nucleus, on the contrary, it has far greater density and a certain stabiHty and is therefore distinct from the cvto- plasm, not so much on chemical as on structural grounds. § 3. Chloroplasts a. Microscopic Structure of the Chloroplasts According to the handbooks of Schurhoff (1924, p. 57), Guil- LiERMOND, Mangenot et Plantefol (1933, p. 1 5 8), Sharp (1934) and KiJSTER (1935a, p. 288), the chloroplasts are microscopically homo- geneous. They are described as hydrogels and both Kuster (1935a) and HoFMEiSTER (1940) even incline to the view that they are in a liquid state of aggregation, though their flattened shape and their autonomic transfiguration (Senn, 1908) would discount this view. As against Kuster's presentation of the matter (1935 a), richly docu- mented as it is, publications have been amassing since 1935 arguing in favour of a microscopic structure in living chloroplasts (Hubert, 1935, p. 369; DouTRELiGNE, 1935; Heitz, 1936a, b; Frey- Wyssling, 1937c; Geitler, 1937; Weier, 1938). All the investigators mentioned find the chloroplasts to be finely granulated and for this reason appeal to Schimper's (1885) and A. Meyer's (1883) grain theory. Schimper's 244 FINE-STRUCTURE OF PROTOPLASM II doctrine states that the chloroplasts consist of a colourless stroma, in which minute granules, lying on the boundary of microscopic visi- bility, are embedded; and these contain the green pigment (Binz, 1892). Colloid research, however, had utterly refuted this view, for the methods employed by colloid optics seemed to show that all living components of the cells are fluid (Kuster, 1935 a, p. 290), optically empty (Guilliermond, 1930) and microscopically homogeneous. Consequently, any kind of microstructure made visible in some way or other was said to be a form of precipitation, structure of coagu- lation, artificial product or artefact. The granular structure of chloro- plasts suffered the same fate. Photographs taken of living cells provided the evidence for the refutation of the theory that the grains in chloroplasts are a product of precipitation. The first microphotographic document may be said to have come from Heitz (1932), who photographed chlorophyll grains next to a living nucleus in the leaf stem of Victoria regia. Doutreligne (1935) considers photography in red light an especially suitable means of proving beyond doubt the inhomogeneous distribution of chloro- phyll in the plastids. Her objects are mosses {Mniuw), Vallisneria, Cahomha and Myriophyllum. Wieler (1936) identifies the grains in a variety of Selaginella. But the most detailed work is undoubtedly that of Heitz (1936a, b), which contains microphotographs of a great number and variety of plants. The grains are decidedly identified in mosses {Physcomitrium^ Hypnum, AInium, Funaria), vascular crypto- gams, very many Monocotyledons and Dicotyledons. Most authors preferred single-layer leaves, such as mosses and fern prothallia, for their observations and Doutreligne avoids even the source of error involved in the use of an embedding medium, using transparent water-plants. Heitz disdains this precaution and includes sections of living tissue in his investigations. One of the things he notices in the leaf of Agapanthus umhellatus is that certain chloroplasts are liable to be damaged (though the cause is not known) and in that state their granular structure is far more clearly apparent than in the undamaged specimens. Evidently this is a kindred case to the fixation of the nuclei, where a barely visible structure in the live state is coarsened in death and the blurred outlines of the optically merging structural components become more sharply defined. Seeing that so many observers have described the plastids as microscopically homogeneous. 3 CHLOROPLASTS 245 we are compelled to assume that the grana are often submicroscopic and only become visible by coarsening. Experience of nuclear struc- tures would seem to imply that, again, it is not a matter of artefacts in this case, but rather of pre-formed structures which, lying below microscopic resolving power, or exhibiting no optically demonstrable phase boundaries, have become visible. The second alternative at the same time shows why the chloroplasts appear to be optically empty in the ultramicroscope (Guilliermond, 1930). Heitz declares that the grana vary in size from 0.5 to 2 /< and that the size is specific to the species. As against this, the granules in light plants are always found to be smaller than in shade plants ; accordingly, the granular size increases from the upper side of the leaf (palisades) towards the underneath (spongy tissue). The grana are especially large and distinct in the chloroplasts of the green fruit of Polygonatum (Menke, 1934a, who, however, calls them artificial products; Weber, 1936). The evidence that the grana are not globules, but platelets, is im- portant (Heitz, 1936b). In the side view of the flat discs of chloro- plasts they look like dense streaks (cf. Fig. 130b, p. 25 5). The Heitz microphotographs reveal no localization of the grana in the periphery; this conflicts with the observations made by Priestley and Irving (1907), ZiRKLE (1926) and Wieler (1936), according to which the colouring matter is accumulated in the cortex and is lacking in the centre. As only the grana contain the pigment, they alone show the fluor- escence of chlorophyll (Heitz, 1936b; Metzner, 1937), appearing bright red, whereas the stroma remains dark. In this way the hetero- geneous distribution of the chlorophyll can be proved indubitably, even in what appears to be optically homogeneous chloroplasts. MoMMAERTS (1938) is of Opinion that the minute green particles occurring in infusions of ground leaves (Noack, 1927) are isolated grana, which he subjects to chemical analysis. Gjr.anick (1938) and Menke (1938b), however, succeeded in obtaining undamaged chloro- plasts from the leaves for chemical examination. Strugger (1950) has discovered that the small amoeboid un- diflerentiated proplastids which exist in dividing meristematic cells already contain a single primary granum. This minute disc multiplies by auto-reproduction. When two grana are formed in this way, the FINE-STRUCTURE OF PROTOPLASM II 246 proplastid divides and each part is provided with one of them. This scheme of multipUcation goes on as long as there is cell division and the number of proplastids increases in the young cells. Only when their definite number has been reached and the cell differentiates do the proplastids evolve to mature plastids. Then the self-reproduction of the grana in the expanding plastid proceeds in a very characteristic way. After splitting parallel to the disc-plane of the granule, the two new platelets remain juxtaposed, split further and pile up, so that cyHnders of grana result with their axes perpendicular to the surface of the flat plastid. It is due to this arrangement that the green colour of the grana is visible in the microscope in spite of their minute thick- ness; in fact, it is not a single granule, but a pile of grana that is observed. The grana produce chlorophyll only in the hght. If they contain but a trace of this pigment, they can easily be discovered in the fluorescence microscope. Before any chlorophyll is synthesized in the proplastids, they must be made visible by staining with rhodamineB. b. Molecular Constituents of Ch lor op lasts Proteins, lipids and the pigments chlorophyll a, chlorophyll b, as also carotene and xanthophyll, which are given the collective name of carotenoids, go to the making of the chloroplasts. Menke finds 47.7% of protein and 37.4% of lipids in the chloroplasts of spinach leaves. They are rich in ash (7.8%) and contain about 7.7% of chloro- phyll (Menke, 1940b). Half the lipids consist of fats, 20% of sterines, TABLE XXIV ANALYSIS OF CHLOROPLASTIC MATTER OF Spinacia ohracea IN % BY WEIGHT (rABINOVITCH, 1 94 5) Menke (1938b) Chibnall (1939) BOT (1939) COMAR (1942) Lipids Protein Ash Residue 37-4 47-7 7.8 7-1 25.1 39.6 16.9 18.4 26-32 42-54 16-25 34 54 7 Chlorophyll .... 7-7 3 CHLOROPLASTS 247 16% of raw wax and 2-7% of phosphatides (Menke and Jacob, T942). Other authors find similar values as shown in Table XXIV (Frey-Wyssling, 1949b). There is no intrinsic chemical difference between the chloroplastic protein and cytoplasmic protein of spinach (Noack and Timm, 1942; TiMM, 1942) ; the former contains a little more histidine and somewhat less lysine and glutamic acid. According to Noack (1930), the cata- lytically active iron (Noack and Liebich, 1941; Liebich, 1941) is bound by adsorption in the stroma. Mommaerts was inclined to view the grana as the containers of the iron, but the grana he used for his Avork were not perfectly pure. Microchemically, the lipid content of the stroma has been definitely proved both by the myelin forms produced by Weber (1933) and Menke (1934a) from chloroplasts, and by the vital staining of the grana by the lipid dye rhodamine B introduced by Strugger (1936/37). The formation of myelin depends upon the following two con- ditions: firstly the lipid molecules must be liberated from any loose linkage to the protein frame so that they can "coalesce"; secondly, they must possess not only lipophiHc, but also hydrophiHc end groups which, as seen on p. 5 6, cause an infiltration of water. The presence ■of water alone does not initiate the emigration of the plastid myelin, from which fact one may infer that the lipids in the chloroplasts have no free hydrophilic groups, but that these are screened ofT, for instance, by the formation of a hpoprotein complex. If, however, they are liberated by saponification in a shghtly alkahne medium (NH4OH), myelin is formed at once. We have fewer exact data on the chemical constitution of the grana. If they do not serve merely as energy traps, but are at the same time the loci of COo assimilation, they must contain proteins in addition to pigments. Euler, Bergman and Hellstrom are of opinion that this system is ten to twenty times the size of a chlorophyll molecule. Mestre (1930) calls the compound between chlorophyll, lipid and protein the "phyllochlorine complex". Stole, borrowing Willstatter's nomenclature (Willstatter and Rohdewald, 1934), called the hypothetic compound "chloroplastin simplex". (It should be noted that in this term the word "plastin" does not cover the sense in which the older authors employed it; they used it to denote the stroma protein, whereas it is here applied to the grana protein.) Stole 248 FINE-STRUCTURE OF PROTOPLASM II and Wiedemann (1941) succeeded in producing this protein con- taining chlorophyll in its pure state. They call the resulting chromo- protein "chloroplastin'". Its molecular weight in the ultracentrifuge was found to be roughly five million. This compound was obtained from thirty different plant species; it shows, as do the haemoglobins of various vertebrata, slight differences, according to the plant species. The chloroplastin of Aspidistra contains about 69% of protein (plastin), 21% of lipids and 8% of pigments, 6% of which, approxi- mately, is chlorophyll. Menke (1940b), finding 7-8% of chlorophyll in toto in the chloroplasts, doubts whether the chloroplastin contains a pure chromoprotein. As, however, the chloroplastin is free from iron, it may nevertheless be assumed that it does not contain all essential constituents of the stroma. We are better informed as to the structure of the pigments in chloro- plasts than on the molecular structure of the protein. One reason for this is that the pigments are easier to isolate, another being that they are of considerable physiological interest. The chlorophyll molecule C55H7205N4Mg is like a tadpole in appearance, having a large head and a long tail (Fig. 1 27). The head consists of four rings of pyrrole linked together to form one porphin ring. This harbours a magnesium atom in the centre and at its periphery are, in chlorophyll <2, four methyl, one ethyl and one vinyl groups and also three oxygenic side chains, viz., one butyric acid, one acetic acid and one formaldehyde residue. The two latter are interconnected laterally (shown by 9 and 10 in Fig. 127); an isocyclic ring is therefore formed, to which has been ascribed the process of assimilation on account of its labile acetic acid-ester configuration (Fischer, 1935 ; Stoll, 1936). The acid groups are esterified with methanol and phytol (CgoHggOH). Chlorophyll b differs from chlorophyll a merely by the substitution of the methyl group at the 3. C atom, shown by a circle in Fig. 127, by a formaldehyde residue -CH = O. There are ten double bonds in the polycyclic ring; they are conjugated, which means to say that they alternate regularly with simple bonds. Systems of conjugated double bonds like this cause absorption of light in short-wave light. Strong absorption in the far red is furthermore induced by the effect of porphin ring formation upon the system of unsaturated bonds. The presence of magnesium only slightly shifts the position of the various absorption bands of this system, but it does affect their intensity. It is there- fore responsible for the green colour of chlorophyll. If the magnesium is removed from the porphin nucleus, the brilliant colouring fades and changes to a dirty olive brown (phaeophorbids). The sUght morphological difference as between chlorophyll h and chlorophyll a suffices to change the bluish CHLOROPLASTS 249 XH-CH3 CH2 \h-CHj CH2 CH, CH2 CH2 o Cm CH-CH^ CH: CH2 II ^ CH I CH2 O I I CH2 -CH I -C< I II '/^CH3 )C-N^ N-Cs^ HCfi Mg rCH C=N N—C. / I \ ICH3 H3C. N,/ H3C HC H2 H2 c-c^ j:h2 -CH, >c^c. ■ C^ II ■CH CH3 o=c- A HC^ M HC^f^y II HC^CH CH, green colour which distinguishes chlorophyll a by increased absorption in the blue, to a yellowish green shade. The head of the chlorophyll molecule has a hydrophilic character owing to the nitrogen atoms of the four pyrrole rings and the co-ordinately bound magnesium. Its long phytol tail, on the other hand, is lipophilic; there is, therefore, in this pigment a clavate molecule with a pronounced lipophilic pole and a lipophobic pole. By contrast, the carotenes, which are unsaturated hydro- carbons of the empiric formula C40H56, are completely lipo- phiHc. The xanthophylls, on the contrary, may contain as many as six hydroxyl groups and are therefore not so decid- edly hydrophobic. Whereas it was formerly held that the carotenoids are dissolved in the chloroplastic lipids, Menke (1940c) is of opinion that, like chlorophyll, they combine with protein molecules to form chromoproteins. As far as carotene is concerned, an argument against this hypo- thesis is provided by the fact that, unUke xanthophyll and chlorophyll, this pigment can ch3Ch.^ be extracted by benzene and other organic solvents from dry leaf powder without any preliminary chemical action. The structural formula of the ^-carotene contained in the leaves is represented in Fig. 127. It is a chain of conjugated double bonds which cause the blue absorption and, therefore, the yellow-to-orange colour; it has me- thylic lateral groups and two terminal rings of six members. The consti- tution of /3-carotene is a matter of paramount importance to vitamin research, for the break-down of the double bond occupying a middle position in the chain (shown by -> in Fig. 127) and addition of water produce two molecules of vitamin A (Karrer, 193 5). ^-carotene is optically inactive. In the a-carotene in carrot roots and in palm oil the double bond between the C atoms marked in Fig. 127 5 and 6 in one of the six-membered WCv ,H ■CH II o HC II HC^QI^ II CH3 "\/y/^V H3CS CH3 AH3 CH3] CH=CH2 " HC^ hI \ yC=C H3C^ ^c-c^ Chlorophyll a fi- H2 H2 Carotene Fig. 127. Molecular structure of the pigments in the chloroplast (it should be noted that it is not certain whether the carotenes possess the trans- configuration or the cis-configuration drawn here) . 250 FINE-STRUCTURE OF PROTOPLASM II rings is shifted to a place between atoms 4 and 5 ; as a result, the C atom marked 6 becomes asymmetrical and the molecule optically active. In the case of y-carotene the six-membered ring is open, the bond between C atoms I and 6 lacking. Small to larger quantities of a- and y-carotene are often present in leaves, as, for example, a-carotene in the leaf of Daucus Carota (Mackinney and Milner, 1933) and y-carotene in Cuscuta salina (Spoehr, 1935, p. 193). To these three carotenes may be added lycopene and others, all of which are distinct from each other by virtue of their melting points and absorption spectra (Smith, 1936). Like ^-carotene, a-carotene and y-carotene are provitamins for the growth factor A, but they produce only half its effect. This is because the two symmetrical halves of ^-carotene have exactly the same chemical constitution as vitamin A, whereas, owing to the slight morphological changes to one of the terminal six-membered rings of a- and y-carotene, only the unchanged half of the structural formula can produce vitamin molecules. With lycopene both the terminal six-membered rings are open, which is why this carotenoid, known chiefly in tomato, has no vitamin A activity at all (Karrer, 1935; Kuhn, 1937). This illustrates most aptly the powerful influence of the special morphology of the molecules upon the specific reactions in the organism. There are also numerous yellow xanthophylls C4 3H56-q(OH)jj. Except for the introduction of OH groups at certain places in the structural formula, their molecules are built up in the same way as the orange carotenes. Crypto- xanthin possesses one of these hydroxyl groups at the C atom marked 3, whereas in the zeaxanthin from the grains of maize both six-membered rings are substituted in this way. There are small amounts of both compounds in leaf xanthophyll, though it mainly consists of another xanthophyll with two OH groups viz., lutein, which has been known for some time from egg yolk. It comprises 50-60% of the xanthophyll (Spoehr, 1935) in the leaves of spinach, gourd, sunflower, lettuce, barley and other leaves. The OH groups cause the beginning of light absorption to shift somewhat towards the shorter wavelengths as compared to ^S-carotene. In carotenoids with three and more oxygen atoms, epioxide-bridges have been discovered (Karrer, 1946). According to the foregoing considerations, the fundamental principle of the molecular structure of all carotenoids is a relatively short chain of un- saturated hydrocarbon with conjugated double bonds. Minor variations in this type of structure give rise to the numerous carotenoids and hydroxyl substitution produces the various xanthophylls (Smith, 1937). As opposed to this variability on the part of the yellow pigments, in higher plants we have the two green pigments, chlorophyll a and h, with their strikingly unvarying constitution. Thanks to this, the percentage of the two chlorophyll pigments contained in leaves can be determined by the quantitative method of spectral analysis (Heierle, 1935 ; Sprecher, Heier- LE and Almasi, 1935). The yellow leaf pigments lend themselves to such analysis only if they are composed of ^-carotene and lutein and nothing 3 CHLOROPLASTS 25I else. By this method Heierle (1935) finds for Amersfoort tobacco at the end of July, for instance, per square metre of leaf surface: chlorophyll a 147.5 n^g' chlorophyll ^53.8 mg, carotene 37.2 mg and xanthophyll 17.8 mg. This represents the famiUar molecular ratio of 3 : i for the green pig- ments and roughly one molecule of carotenoids to every two chlorophyll molecules (about 1/3 molecule of carotene and 2/3 molecule of xanthophyll). By means of chromatographic adsorption Seybold (1941) made comparative measurements and found that the molar ratios just given do not invariably exist between the pigments. Chlorophyll h, for instance, may be present in far smaller quantities, or may not occur at ail, this applying notably to certain algae (Seybold, Egle and Hulsbruch, 1941). Instead, those groups of algae may contain other varieties of the green pigment, such as chlorophyll c or chlorophyll (/(Aronoff, 1950). c. Suhmicroscopic Structure of the Chloroplasts State of chlorophyll in the chloroplast. Granular chlorophyll and mole- cular chlorophyll solutions (in acetone, alcohol, etc.; Fig. 128a) show red fluorescence when exposed to light rays ; the fluorescence is pro- AL (OhJs c Fig. 128. Chlorophvll molecule, a) Molecular dispersion; b) colloid particle; ) p //!■.■'■-•.■.■-■.■.•.•.■■."■■■■ ^) Fig. 130. Structure of chloroplasts. a) Submicroscopic layer structur neglecting the grana structure. P protein layer, L lipid layer, with indication of the optical character (from Frey- Wyssling, 1937O; ^) scheme of a cross-section of a chloroplast in ultraviolet light (from Menke, 1940^). 256 FINE-STRUCTURE OF PROTOPLASM II some scepticism. However, gold-stained chloroplasts in profile clearly exhibited dichroism (cf. p. 84, loi), which is indicative of a laminar texture (Menke and Kuster, 1938). Further proof of the lamellar texture was provided by the large chloroplasts oi Anthoceros, that classical object which, at the instigation of Ernst, had already been appealed to so fruitfully in the dispute over the relationship between plastids and chondriosomes (Scherrer, 1 9 14). Menke and Koydl (1939) identified layers at the limit of microscopic resolution in cross-sections through the chloroplasts of Anthoceros using the enhanced resolving power of the UV microscope. Not only do the big chloroplasts without grana of Anthoceros and the Conjugatae algae appear to be laminated, but also the granulated chloroplasts {Selaginella, Phaseolus). The grana are united by thin lamellae, which induced Menke (i94od) to devise the plan of Fig. 130b of a section through the discoid chloroplasts of the higher plants. The pile-like arrangement of the grana (Strugger, 1950, 195 1) is clearly visible. Electron microscopy (Kausche and Rusk a, 1940; Menke, 1940 a; Algera, Beyer, v. Iterson, Karstens and Thung, 1947; Granick and Porter, 1947). Besides stroma and grana, a distinct boundary layer has been disclosed (Frey-Wyssling and Muhlethaler, 1949a) as a third morphological element of the chloroplast (Fig. 1 3 la, p. 259). This layer must consist essentially of proteins, as it displays the properties of a solid and does not show any sign of the liquid or semi- liquid state characteristic of lipid matter. It is probable that the living boundary contains lipids, but their amount must be small as compared with the total lipid mass in the chloroplast. Obviously they join the emigrating myelin. The proteins of this plastid layer must be of the fibrous type ; otherwise the formation of a membrane would not be possible when dried. The formation of strands of stretched chloro- plasts (Kuster, 1935c) is probably due to this protein. How much the dried membrane differs from its natural state in the living chloro- plast is not known. Under the membrane, the grana are visible as discs. The stroma, on the other hand, does not show any conspicuous structure. Wyckoff (1949) has given evidence of globular macromolecules about 250- 300 A in diameter, which lie on and between the grana. If the plastid membrane has burst, as usually occurs during the preparation of the 3 CHLOROPLASTS 257 chloroplasts, the whole carrier film is sprinkled with these globular bodies (Fig. 131c, d). This behaviour would indicate that the stroma is a corpuscular dispersion of macromolecules, i.e., a sol. Since a sol has no framework, the characteristic shape of the chloroplast must be due to its membrane, much the same as in erythrocytes, and to its internal lamination (Fig. 130b). The chloroplast can change its shape (Senn, 1908), or even form processes (Heitz, 1952); this faculty must also be ascribed to the membrane, which may be compared with the ectoplasm of creeping protozoa. This again argues in favour of a protein rather than of a lipid ground mass of the plastid membrane. The grana supporting lamellae suggested by Menke (Fig. 130b) and Strugger (195 i) have not yet been found in the electron micro- scope. We may ask whether the macromolecules found by Wyckoff (1949) represent lipo-proteins or only proteins. It is almost certain that the latter is the case. The preparations show very thin flat discs (Kausche and Ruska, 1940) of various diameters up to 5 // and only 100-200 A thick. It can be shown that before desiccation these discs wxre in a semi-liquid state. They never have folds, as the plastid membrane has, and dry perfectly smoothly on the carrier film, even if they include isolated grana (Fig. 131c). There has been much dis- cussion on the nature of these discs. They have been looked upon as protein lamellae (Menke, 1940a) or phosphatide bladders (Algera and co-workers, 1947) (which is unlikely, as the chloroplast contains only 0.5-2.5% of phosphatides), but there is no doubt that they represent the total lipid matter of the chloroplast and must be con- sidered as myelin forms. Fig. 131c shows how this myelin flows out of a fraction of a disintegrated chloroplast. It is likely that the grana lipids have also emigrated, because, as seen in the electron microscope, the grana consist of proteins only. Washing with lipo-solvents does not alter them (Menke, 1940 a; Granick and Porter, 1947). They seem to be layered Hke a low pile of coins. Occasionally such a pile appears to be overturned (Fig. 13 id), when a number of very thin lamellae, all of the same diameter, are visible. The question is justified, whether these lamellae are really lamellar parts of grana or perhaps ghosts of whole grana. However, Steinmann has disclosed in Aspidistra chloroplasts as many as 30 of these lamellae in the same pile (unpublished). This rules out any 258 FINE-STRUCTURE OF PROTOPLASM 11 errors in interpretation, since a pile of grana consists of only about 8 microscopic discs (Strugger, 195 i). The submicroscopic lamellae must consist of protein. In the living^ state, the lipids in the grana were probably located between these protein layers. If this picture can be substantiated by further research^ the grana of the chloroplast would represent a layered composite body with alternating protein and hpid lamellae. The chlorophyll is closely associated with the grana lipids, because it emigrates together with them; on the other hand, Menke (1938c, 1943) points out that chloro- phyll migrating with the Hpids imparts conspicuous dichroism to the myelin tubes, lacking in the profile of the chloroplast. Hence a simple combination of hpids and chlorophyll is excluded, which is a further argument in favour of the existence of a chromoprotein. From Fig. 13 id it may be concluded that this chromoprotein is arranged in layers. If this conception of the arrangement of the chromoprotein be correct, the principle of laminar surface develop- ment can be consistently pursued from the molecular to the macro- scopical region. The molecular layers compose the discoid, sub- microscopic to microscopic grana (Fig. 130a, p. 255); these, again, lie in layers in discoid or laminar chloroplasts and finally the chloro- phyll grana are exposed to the light, again in foliar laminae. Tracing thus a given morphological principle through several orders of magnitude, we are provided with an interesting counterpart to fibre structure, in which linear development plays a similar part. The laminar series: molecular layer/grana/chloroplast/foliar laminae may be compared with the linear series : chain molecule/microfibril/fibre/ fibre bundles of the pericycle. It should be emphasized that in both cases the form birefringence has been the key to the submicro- scopic structure, viz., the discovery of the rodlet birefringence in fibres and of the platelet birefringence in chloroplasts. Chloroplastin and the unit of assimilation. The definite estabHshment of the grana as the only loci in the chloroplast containing chlorophyll, calls for a discussion concerning the biochemical concept of chloro- plastin. There is no doubt that the grana represent a high concen- tration of chlorophyll. According to Granick (1949), the chloroplast of spinach contains only 40-60 grana, 0.6 ^ in diameter and 0.08 fjt thick. Since in some instances it has been possible to photograph the grana in profile with the fight microscope (Heitz, 1932), this sub- CHLOROPLASTS 259 Fig. : 3 1 . Chloruplasts ot tobacctj leaves in the electron microscope (t rom !■ rey-\\ yssling and MunLETHALER, 1949a). a) Chloroplast membrane; b) grana; c) myelin covering grana and globular macromolecules of disintegrated stroma; d) intact granum and layers of an overturned granum. 26o FINE-STRUCTURE OF PROTOPLASM II microscopic thickness of 0.08 fx may be due to desiccation during the preparation for the electron microscope, and we may estimate the thickness of the fresh grana to be about o.i 5 [x. The whole chloroplast has a diameter of 5 fx and its thickness in the fresh state is about half of this. If we calculate the volume of the chloroplast as an ellipsoid, 4/3 X 2.5^ X 1.25 X n, and that of the 50 grana as cylindrical discs, 50 X 0.3^ X 0.15 X n, wt obtain a volume ratio of 15/1. Thus, the total volume of the grana would be only 1/ 1 5 or 7 % of that of the whole chloroplast. Since there is 7.7% of chlorophyll in the chloro- plast (Menke, 1940b), this would mean that the grana consist entirely of chlorophyll. This is obviously impossible, for the grana are still visible in the electron microscope when the pigments are extracted. We must conclude, therefore, that the discs visible in Fig. 131b do not represent individual grana, but piles of grana. The number of piles in the chloroplast of tobacco leaves is about 50, thus the same as in spinach leaves, and their diameter 0.4 fx. The chloroplast of about 2.5 // thickness can lodge not more than 12 layers of grana. With these figures the volume ratio chloroplast/total number of grana is 3, i.e. the grana occupy 1/3 and the stroma 2/3 of the plastid volume^. This ratio enables us to calculate the chlorophyll content of the grana protein. According to Table XXIV (p. 246), half the weight of the chloroplast is protein and 7.7% chlorophyll (mol.wt. 893). This yields a molar xatio of 3 chlorophyll to i Svedberg unit (mol.wt. 17600). Since the chlorophyll is restricted to the grana and their volume being only one third of the chloroplast, this ratio must be 9/1 in the grana, if the protein concentration is the same as in the stroma. This result seems to prove that chlorophyll cannot be a prosthetic group of an enzyme, for, considered as a co-enzyme, its carrier would have a molecular weight as low as 2000, which has never been found for apo-enzymes. Stoll's chloroplastin (1936) has a molecular weight of roughly 5 millions. If it is really the chromoprotein of the chloroplast, it must come from the grana alone and cannot be contaminated by stroma protein. It is doubtful whether these two proteins can be separated quantitatively by fractionated precipitation. Supposing the chloro- 1 Thus about 25 % by weight of the grana consists of chlorophyll; this is astonishingly high, as compared with the haematin (mol. wt. 592) content of the erythrocytes (p. 265) which is only ca. 3 % of the cell interior. 3 CHLOROPLASTS 261 plastin with the molecular weight 5 millions to be really the chromo- protein of the chloroplast, it ought to hold 2500 chlorophyll molecules, but in reality it contains only about 420. This indicates that the chloro- plastin is a mixture with stroma constituents rather than a pure com- pound from the grana. On the other hand, physiologists find that a number of chlorophyll molecules as large as calculated above is necessary for the assimilation of one molecule of COg. That number is called unit of assimilation. Whereas chemists think of the photosynthetic process as associated with the chlorophyll molecule (Stoll, 1936), physiologists tend rather to regard the pigment merely as an energy trap ajid to attribute the actual chemical action of the gradual hydrogenation to the proteins in the chloroplast (Rabinowitch, 1945). This is inferred partly from Blackman's dark reaction (1905), but mainly from facts established by Emerson and Arnold (1932), according to which a. unit of assimi- lation of roughly 2500 chlorophyll molecules is needed for the re- duction of one CO2 molecule. Gaffron and Wohl (1956) calculate about 1000 molecules for this same unit. This observed fact calls into question all attempts to deduce the mechanism of assimilation from the chemical constitution of the chlorophyll molecule (Willstatter, 1933; Franck, 19.35; Stoll, 1936). Gaffron and Wohl state that the pigment acts merely as a specific energy transmitter and that a very large number of chlorophyll molecules would be required to capture the necessary quanta of light for the assimilation of one CO2 molecule (Warburg and Negelein, 1923; Scplmucker, 1930; Eymers and Wassink, 1938; Emerson and Lewis, 1939). It is to be expected that the occurrence of these units of assimilation will be expressed morpho- logically in some way. Heitz (1936a) presumes that the grana may be involved. This, however, cannot be so, for if, as Euler, Bergman and Hellstrom (1934) state, a chloroplast contains 1.65-10^ chloro- phyll molecules, there would have to be something like 10^ or a million grana, whereas the actual number is about 600. In a bi- molecular layer, 2000 chlorophyll molecules would occupy a surface of 1000 X 225 A^ =^ z.i^ X io~^ fj,^. As a square, this surface has a side of only 0.05 /u. Therefore the unit of assimilation is certainly amicroscopical. Seeing that a chloroplastin macromolecule in the grana ought to contain about 2500 chlorophvll molecules, the question naturally 262 FINE-STRUCTURE OF PROTOPLASM II arises whether the unit of assimilation is identical with the chloro- plastin unit. This would simplify our terminology. But as long as it cannot be proved that chloroplastin derives from the grana alone, the coincidence of the number of chlorophyll molecules in the unit of assimilation and in the chloroplastin molecule seems to be only incidental. § 4. Erythrocytes a. The Microscopic Structure of Erythrocytes It is not only their lack of a nucleus which makes the red blood corpuscles of mammals a cytological curiosity, but it is also the peculiar shape of the cell. Seen from the top in the microscope, they look like round discs, the bound- ary of the cross-section of which is curiously sinuate, instead of being planoparallel. Thus the erythrocytes are biconcave discs. This remarkable shape of the Fig. 1^2. Cross-section of the red cell of man. . . • j j. i j ,^ „ , . , 1/ / J , rv cross-section is said to be due ab = 8.55/i; thickness li (cd 4- ef) = 2.40//; thickness gh = 1.02 /i (from Ponder, 1934)- to the function of the red blood corpuscles, since from a surface thus shaped the interior of the cell can be easily supplied with oxygen by diffusion, whereas a globular shape would entail greater poverty of oxygen in the centre than in the surface layers and, with a piano- parallel disc, the edge would be richer in oxygen than the centre. The discs remain biconvex in shape as long as the erythrocytes are suspended in the blood plasma or in serum, but they round up directly if the medium is changed by the addition of lecithin to the blood plasma. It is a remarkable fact that the same thing happens when a thin layer of them is covered with a cover glass. Ponder (1934), discussing many possible causes of this phenomenon, omits to mention the change in r^ of the medium and asphyxiation, which all living cells undergo after some time in the thin layer under the cover glass. Under certain circumstances rounded blood corpuscles can be restored to their initial biconcave disc shape by the addition of serum. As any experiment with erythrocytes involves possible transfor- mation, it is not an easy task to establish their true cross-sectional 4 ERYTHROCYTES 263 shape. Ponder (1954) obtained the image shown in Fig. 132 by a series of microphotographs with an objective of the least possible focal depth. With retention of the volume, the transformation to spheres is effected by surface changes only. For instance, the biconcave erythrocyte of the rabbit has a surface of no fi^, whereas that of the globular form is only 70 fi^ (reduction in surface of 36%). A further indication of surface changeability is provided by the dented blood corpuscles, which are transitions between the biconcave discs and the globules, or the curious thorn-apple forms which arise vmder certain conditison. These facts make it plain that surface forces are responsible for the shape of the erythrocytes. Gough (1924) points out that surface-enlarging forces must be active in the erythrocytes, conducive to expansion of the surface of contact with the suspension liquid, as in the case of the myelin forms. The largest surface would be obtained if the blood corpuscle were flattened to the thinnest possible disc. On the other hand, there is some shght surface tension in the blood corpuscles (presumably of the order of i dyne/cm, cf. Table XXI, p. 166), tending to reduce the surface and to round off the erythrocytes if other factors do not interfere. Now it mav be that the exceptional shape of the erythrocytes of mammals represents some kind of equilibrium between the surface- enlarging and surface-reducing forces. If that be so, the membrane of the erythrocytes should have the properties of a mesophase. No form of equilibrium can, however, be mathematically computed from the cross-section in Fig. 132 and Ponder (1934, p. 89) therefore inclines to the belief that there must be a certain amount of internal solidity. The micrurgical investigations of Seifriz (1927, 1929) tend to endorse this, for they show that deformed and elongated erythrocytes have some slight elasticity. Dervichian does not agree with this view (Dervichian, Fournet and Guinier; 1947). Inner structure. The various theories as to the internal structure of the red blood corpuscles are expressed in the following two views. One school regards the anucleate erythrocytes as enclosed in a mem- brane which gradually changes towards the interior into a very loosely knit spongy structure, in which the red blood pigment is embedded. Some support for this view is afforded by the network structure which can be made visible in young erythrocytes by suitable fixation and staining. Representatives of the other school of thought, however. 264 FINE-STRUCTURE OF PROTOPLASM II dismiss this network as mere artefact. To their way of thinking, the erythrocyte consists merely of a balloon-like membrane, a view which has some backing through the absence of any microscopic structure in the Hving cell interior as seen in the ultra microscope or illuminated by ultraviolet rays. This view is also shared by most of the research workers who have studied haemolysis. For, if the erythrocytes are damaged mechanically, either by heat or freezing, or by immersion in sufficiently hypotonic or hypertonic solutions, the contents of the cell extravasate with the red blood pigment and a colourless sheath re- mains, which is called ghost, or the stroma. These facts notwithstanding, the contents of the erythrocytes are not to be considered as a sol-like liquid of no organized intrinsic structure, an error committed by the older investigators and, more recently, by Gough (1924'). The relative viscosity of the cell contents is 30 (see Table XXII, p. 169) and Ponder (1934) states that the interior of the cell shows respiration like other cells. Although the erythrocyte membrane has been proved to contain all the chemically identifiable substances of the blood corpuscles with the exception of the blood pigment and the salt content, the assumption clearly must be that the contents of the cell, far from being an unorganized liquid, is a partially gelated cytoplasm, the organization of which is easily destroyed when damage is inflicted. The thickness of the ghost membrane has been measured by numerous investigators with a wide variety of results ranging from 15 to 700 vafx (Jung, 1950). This seems rather embarrassing. But when the methods used for the measurements are considered the results can be classified into two groups, viz. those obtained from dried ghosts (electron microscope. Fig. 135, p. 272, Wolpers, 1941; leptoscope^ Waugh, 1950), yielding 15-25 m^, and those from hydrated ghosts (dark field observation, Lepeschkin, 1927; micrurgy, Seifriz, 1927) with about 500 m/<. The last figure has also been found by Mitchison (1950b), who has thrown down the ghosts by a centrifugal force of 1 10,000 g to a compact mass which is still 5 5 % of the volume of the intact red cell. From this result it follows that the swollen membrane is as thick as half the depth of the erythrocyte (diameter c-d of Fig. 132) and that it shrinks when dried to 1/25 of this size! The inner part of the membrane, therefore, represents in vivo a very loose gel with only 4% protein, which fills almost the whole erythrocyte. 4 • ERYTHROCYTES 265 b. Molecular Constituents of the Erythrocytes Erythrocytes consist approximately of two-thirds water and one- third dry residue, which is mainly composed of the red blood pig- ment, haemoglobin, and salts. It is interesting to note that potassium predominates over sodium as cation of the salts. Small amounts of protein foreign to haemoglobin and of lipids constitute the ery- throcyte membrane. Haemoglobin. The red blood pigment is a chromoprotein, Hke chloroplastin in green leaves; yet the Hnk between chromogen and protein is closer than in chlorophyll and the blood pigment therefore emerges as protein from the stroma in haemolysis. Haemochromogen is a labile porphyrin compound which, outside the organism, is transformed into the more stable haematin. The composition of this compound is C34H3204N4FeCl and it is closely akin to chlorophyll (Gr.\nick, 1948). The main differences are that in the centre of the porphin ring there is, instead of magnesium, tri- valent iron, the third valency of which imparts a saHne nature to the compound usually neutralized by the anion chlorine; and the absence both of the phytol chain and the iso-cyclic ring of the C atoms 6-9-10 (see Fig. 127, p. 249). As a result of the missing phytol chain the haematin appears to be morphologically more compressed and less markedly polar than chlorophyll. The protein carrier, to which the haemochromogen is attached is called "globin". The haemoglobin molecule is of a thickset rod-like shape with 57 A diameter and 34 A height (Perutz, 1948). On the basis of the iron content its molecular weight is computed at 16,000 to 17,000 (K.A.RRER, 1941), while the reading in the uhracentrifuge is 69,000, i.e., about four times the value (Svedberg's law of multiples, see p. 141). Stromatin. Jorpes (1932) states that approximately 4% of the total protein content of the erythrocytes consists of a protein foreign to haemoglobin, which is contained in the erythrocyte sheath and is therefore described as stromatin. According to Winkler and Bungen- BERG DE Jong (1941), its I.E.P. is at p^ 5.2. Analysis of the hae- molyzed membrane of erythrocytes shows that there is 80 °o of stromatin and 20 °o of lipids. Phospholipids. The bulk of the hpids consist of phosphatides, notably lecithin (Fig. 93, p. 138), besides which there are insignificant amounts of cephaiin and sphingomyelin. The I.E.P. of the phospho- 266 FINE-STRUCTURE OF PROTOPLASM II lipids is at p^ 2.7. They are thought to play a decisive part in the permeability phenomena of the red blood corpuscles. Cholesterol. Approximately one molecule of cholesterol is found for every four phosphatide molecules in the stroma (exact ratio 3.5:1, Winkler and Bungenberg de Jong, 1941). As may be seen in Fig. 92 (p. 138), cholesterol, unlike the phosphatides, possesses no ionogenic groups. Bungenberg de Jong therefore assigns to it an important part in the formation and build-up of lipid structures, for,- in a lecithin solution, the individual lipid molecules remain separated from each other as the result of their negative charge. Although the fatty acid chains have a tendency to agglomerate, the repellent effect of the ionized phosphoric acid groups predominates and the molecules are therefore kept at a distance from each other. If cholesterol is added to a solution of this kind, these neutral mole- cules are able to penetrate in between the lecithin molecules and association follows as the result of Van der Waals cohesive forces, as the repelling action of the charges does not span the width of the cholesterol molecule. Cholesterol therefore acts as a sensitizer in the precipitation of lipid solutions with ionogenic groups. Conversely, in lipid films of phosphatides, cholesterol acts as a stabiliser, as it counter- acts solution of the film by ionogenic influences. Nucleic acids are only present during the development of the ery- throcytes in the bone marrow. The stem cells contain 5 % cytoplasmic nucleic acid, but during differentiation and maturation of the red cell, its concentration drops to below 0.5% (Thorell, 1948). c. Submicroscopic Structure of Erythrocytes Stromatin as tricompkx system. Winkler and Bungenberg de Jong (1941) have pubhshed an instructive design of the structure of the erythrocyte sheath (Fig. 133). By exact measurement of the electric migration velocity of the red blood corpuscles in the most various salt solutions, these investigators find quantitatively the same be- haviour as in phosphatides, from which they conclude that the surface of the erythrocytes is covered by a phosphatide film (layer I in Fig. 133), which is stabilized by cholesterol. The I.E. P. of the stroma with Ph 5.2 being between that of the phospholipids (2.7) and of the stromatin (5.8), it is assumed that the phosphoHpids form a complex system with the stromatin (layer IV), their positive choline groups ERYTHROCYTES 267 entering into relationship with the anionic end groups of the protein (layer III). Haemolysis experiments have further shown how calcium ions consolidate the erythrocyte membrane and stabiHze it. In layer II the calcium ions, with their strong positive charge, are therefore allocated between the negative phosphoric acid groups of the lecithin and a more powerful ionogenic I t YL 280i i)^i^i^i^»C c qF M E mn-i^-rom^i^Hi / 20& cohesion is thereby attained. Thus the stroma is regarded as a com- plex system consisting of phos- phatide-calcium ions, stromatin protein, and the regular distri- bution of charge brings with it a definite arrangement and orien- tation of the various components m ^2o^\ of the system. The tricomplex sys- tem is completed by an assumed complex linkage of the haemo- globin (layer VI) with anionic end groups in layer V to cationic groups of the stromatin. The design of Fig. 1 3 3 is further complicated by layer A. This re- presents an incomplete film of polar lipids, which turn their lipo- philic side towards the monomo- lecular phosphatide layer I and their hydrophilic pole outwards (fat, fatty acids, possibly chole- sterol). It is necessary to assume this, for, without the layer A, the erythrocytes would agglutinate in aqueous solutions and, when shaken out with paraffin oil, would pass over into the lipid phase, neither of which they do. The scheme devised by Winkler and Bungenberg de Jong (1941) explains manv properties of erythrocytes, e.g., it makes allowance for the lipid filter theory of permeability, there being a fipid film with molecular pores (where the cholesterin covering is lacking). It ex- plains the permeability to anions which is characteristic of erythro- cvtes, as the calcium ion layer III debars the cations. The same layer Fig. 133. Molecular structure of the en- velope of the red cell from Winkler and Bungenberg DE Jong (i940-'4i); • anionic groups ; o cationic groups or cations (Ca) ; shaded: cholesterin; ^ phospholipid acid; ch cholesterin ester; • fatty acid. 268 FINE-STRUCTURE OF PROTOPLASM II of ions, with its water of hydration, is responsible for the effect of hydrating and dehydrating ions upon the properties of the erythro- cytes. According to Fricke (1925), the electric properties of the wall of the erythrocytes are such that the existence must be assumed of a non-conductive layer 3 3 A thick. This thickness corresponds to the lipidic part of the phosphatide layer I. Gorter and Grendel (1925) assume that there is a bimolecular lipid film on the basis of the lipid content of the erythrocytes; and this claim is likewise partly met. Finally, Winkler and Bungenberg de Jong calculate from the stromatin and lipid contents of the erythrocytes of pigs (19.2, or 3.5 mg per ml of blood) that the orientated lipid molecules just cover the surface of the blood corpuscles in the manner indicated (Fig. 133) and that the layer of stromatin below is 1 20 A thick. From this we get 1 50 A as the thickness of the total erythrocyte membfane (without layer A) which, surprisingly, is of about the same order of magnitude as the data obtained by Wolpers (1941) by means of electron optics. However, this is only incidental, since Fig. 133 does not refer to the dried, but to the hydrated envelope. Although this explanation of many interesting phenomena as- sociated with the morphology and physiology of erythrocytes is un- disputed, the model of Fig. 133 still raises a number of difficulties. One of the first points to be noted is that analysis of the erythrocytes has not revealed the presence of calcium. True, Winkler and Bungenberg de Jong have calculated that the quantity of Ca present is so small that it would escape detection in analysis, but they never- theless consider it improbable that, given the percentage of calcium in the blood serum, no Ca ions should be adsorbed on the erythrocyte membrane. In the transition from the biconcave disc shape to the globular, the surface must shrink by 37%. It is not clear how this could take place without causing change of structure since, compared to their normal distances, the molecules are already densely packed. An argument against the parallel radial orientation of all the molecules is the slight optical anisotropy of the erythrocytes. Stro- matin and haemoglobin can scarcely be said to represent chain mole- cules; on the contrary, haemoglobin is known to be a globular molecule. Should stromatin be filamentous, it would seem to me that the orientation of those threads, c^iven their great length, is more likely to be parallel to the surface than a radial one, as suggested. ERYTHROCYTES 269 Winkler and Bungenberg de Jong discuss this possibility; but, iinding that the number of anionic COOH groups of the side chains is not large enough for their tricomplex system, they place the main chains perpendicular to the surface of the cell. Haemoglobin as a solute in close packing. Although the concentration of haemoglobin reaches 34% in the red cell, it does not crystallize; Fig. 134. Close packing of haemoglobin in the erythrocyte (from Jung, 1950). t diameter, h height, d body diagonal of the haemoglobin molecule, a distance of molecular layers. The size of a hydrated Li + and K+ ion is given for comparison. it iills the erythrocyte as an isotropic solute. On the other hand, an X-ray period of 62 A is furnished by living cells (Dervichian, FouBusiET and Guinier, 1947). This period can be explained as follows (Jung, 1950): The haemoglobin molecules are covered by a hydration layer of 3 A, so that the dimensions of its thickset cylinder are 63 A for the diameter /, 40 A for the height h and 74. 5 A for the body diagonal d. If these molecules are allowed free rotation, every one requires a spherical space of 74.5 A diameter (Fig. 134). Further, if these spheres are arranged in hexagonal closest packing, a layer distance of 61 A results, which is consistent with the X-ray period found. Therefore, the state of the haemoglobin in the erythrocyte is that of the densest solution possible, whose concentration has been calculated to be 34*^0. It is evident that such an arrangement is most favourable for the gas exchange of O, and COo. But why is it that such a saturated ZyO FINE-STRUCTURE OF PROTOPLASM II solution does not crystallize? As a matter of fact, every disturbance of the existing equilibrium, say by a hypertonic salt solution or by formation of sickle-shaped cells in anaemic venous blood (Perutz and MiTCHisoN, 1950), provokes the crystallization which is re- cognized by the birefringence of the hitherto isotropic haemoglobin. The possibility exists that in the swollen erythrocyte traces of stro- matin between groups of haemoglobin molecules prevent the crystal- lization which occurs as soon as this stabiHzing system is destroyed. Birefringence. Fresh problems arise as soon the optics of erythrocytes is taken into consideration. Rabbit's red cells, carefully haemolyzed by freezing and thawing, are birefringent (Schmitt,Bear and Ponder 1936, 1938), exhibiting a very faint polarization cross. With respect to the cell radius, the birefringence is slightly negative in isotonic salt solution, but positive polarization crosses are clearly visible in glycerol mixtures. The inference from imbibition tests of this kind is that, as in the case of the chloroplasts, in the sheaths of the erythrocytes there is positive intrinsic birefringence of the embedded lipids, upon which is imposed a negative form birefringence of the protein frame- work. Lipid solvents, such as butyl and amyl alcohol, produce distinctly negative polarization crosses, abolishing the intrinsic bire- fringence of the lipids and bringing the negative form birefringence out clearly. ScHMiTT, Bear and Ponder come to the conclusion that there must be a composite body with alternating protein and lipid lamellae. The lipid layers, they think, must be bimolecular on account of the hydrophiHc bias of the stromatin. This view conflicts with the calculations made by Gorter and Grendel (i92 5)> according to which the lipid content of the erythrocytes would be just sufficient for a single bimolecular covering. The possible layering throughout the stroma would only be lipid-protein-cavity-protein-lipid. Conse- quently, unless those authors' statements are incorrect, it is difficult to see how there can be a composite body of protein and lipid, like that proved for the chloroplasts. Another possible explanation, taking the observed facts into ac- count, is that the stromatin is loosely layered and is in itself a Wiener composite body. In this case, too, the positive intrinsic birefringence of the hpid skin overlays the negative form birefringence, the problem, however, still being whether the lipid birefringence would then be 4 ERYTHROCYTES 27I perceptible at all. The probable retardation F can be calculated with the aid of the formula on p. 86 by inserting the value o.oi i for the birefringence An, which Bear and Schmitt (1936) set down for orientated lipid in the nerve sheath. In rabbits, the diameter d^ of the, supposedly, hollow cylindrical rim of erythrocytes is i.j /n (cf. c-d in Fig. 132, p. 262), and d, is shorter by twice the thickness of the bi- molecular Hpid layer (4 X 3 m//), i.e., 1.688 //. The value for the retardation F is then a httle above 1.8 m^. This is a value which, though at the lower limit of quantitative mensurability with sensitive compensators, may, by suitable polarizing optics, be revealed quali- tatively. This shows that a single bimolecular lipid layer suffices to produce the faint positive intrinsic birefringence detected by Schmitt, Bear and Ponder. Both the quantity of lipid present and the slight intrinsic bire- frino-ence witness to the fact that there can hardly be more than a double film of orientated lipid molecules in the erythrocyte. This eliminates the possibility of a protein-hpid layer composite body, such as demonstrated in chloroplasts. To account for the lamellar birefringence, therefore, one is forced to assume that the stromatin protein is lamellar with, maybe, layers of water in between. These need not necessarily be continuous; indeed, they are more probably cavities shaped somewhat Hke lenses (Fig. 136, p. 272). On this as- sumption the direction in which the stromatin molecules of Fig. 135 (p. 267) (layer IV) are orientated must undoubtedly be turned through an angle of 90° and lie parallel to the erythrocyte surface. MiTCHisoN (1950b) is of opinion that the small amount of hpids cannot contribute anything to the birefringence of the erythrocyte. According to him, the birefringence of a bimolecular lipid layer 6 m/t thick is not measurable, due to diffraction errors. He attributes both the negative form and the positive intrinsic double refraction to the stromatin by assuming that radially oriented looped polypeptide chains are lodged in the submicroscopic stromatin layers. Such an arrangement seems to be rather unlikely. Electron microscopy. Apart from fibres and diatoms, erythrocytes were the first cytological object to produce good and impressive images in the electron microscope (Wolpers, 1941). This is due to their ability to withstand complete drying without any essential change in structure. 272 FINE-STRUCTURE OF PROTOPLASM 11 The photographs of the residue of haemolysis (Fig. 155) merely show a folded membrane. No inner structure is visible, which, ac- cording to WoLPERS, proves the balloon theory of the structure of erythrocytes. The average thickness of the membrane is 25 m/j,. However, this measurement by Wolpers refers to the dried envelope which has been reduced to 1/25 of its thickness in vivo. After suitable extraction of the lipid, Wolpers found the erythro- 1 -. ■» "■ Fig. 135 Fig. 136 Fig. 135. Membrane of red cells, osmotic tixation. Electron microscope 9500:1 (from Wolpers, 1941). Fig. 136. Stretched membrane of red cell in electron microscope. Image scale 51,000:1 (from Wolpers, 1941). cyte membrane to be porous. He therefore discards as improbable the layer structure inferred from observations in the polarization micro- scope. He also rejects the idea of a mosaic structure, which his electron micrographs would at first sight seem to suggest; for he detected a network structure in stretched erythrocyte membranes which had been fixed with osmium tetroxide after extraction of the lipids (Fig. 136). This induces him to believe that the stromatin has ^i frame structure, in the meshes of which he imagines the lipids to be embedded. Whether this opinion is shared, or the meshes are thought to be free from lipids and filled with an aqueous phase, depends upon the re- jection or acceptance of a superficial double film of lipid. However this may be, the optically proved lamellar structure must not be ignored; rather should an attempt be made to reconcile the two findings. ^ ERYTHROCYTES 273 A consistent picture is obtained if the filamentous protein frame is thought of as stratified parallel to the surface and the meshes as shallow, tangentially extended lenses, when the body of the frame- work will exhibit layer birefringence. Under these circumstances, certainly, no pores would be visible in the top view of the skin. The impression received is that sieve-like images are artefacts and not natural structures. This suspicion is strengthened when one examines Jung's photographs (1942) of erythrocyte membranes denatured by heat haemolysis. There are similar sieve images, with even larger pores. More recent electron micrographs of ghosts by Bessis and Bricka (1949) and 2aCek and Rosenberg (1950) do not show any sieve pores, but a coherent fine granulated structure. Without doubt the surface of these membranes is formed by aggregated globular protein mole- cules, which leave only small capillaries between each other. It is open to discussion how the haemoglobin molecules with a diameter of action of 74. 5 A can diffuse across such a membrane with the velocity characteristic for haemolysis. Probably the texture of the hydrated membrane is much looser in vivo than in the completely dried state necessary for the electron microscopic observation. If the capillaries in the membrane appear to be too narrow for haemoglobin, haemo- lysis must locally destroy submicroscopic parts of the membrane where the haemoglobin can freely escape. The electron microscopy does not give evidence of any such mosaic structure of the erythrocyte membrane, which has often been postulated for the understanding of the complicated permeability phenomena (Ponder, 1948). Putting together what we know with fair certainty of the sub- microscopic intrinsic structure of the erythrocyte membrane, we must come to the conclusion that the stromatin has a coherent texture which appears to be laminated, on account of the form birefringence. In the dry state there are lens-shaped or flat submicroscopic spaces. The lipids envelop the whole surface of the erythrocytes in a continuous film. The quantity of lipid is too small for a protein-lipid layer body. In the hydrated state the stromatin is considerably swollen and it is likely that, in vivo, the spacious meshes of this dilute gel are filled with haemoglobin, which assumes the special state of a solute in close packing. This expose of the microstructure of erythrocytes demonstrates impressively the fact that submicroscopic morphology cannot be 274 FINE-STRUCTURE OF PROTOPLASM II inferred from either the indirect methods, or from direct electron microscopy, alone, but that the two modes of enquiry should be complementary and the results obtained with the one should be scrutinized in the light of the data produced by the other. § 5 . Gametes Gametes are very highly differentiated cells with the faculty of transmitting to the zygote the capacity of developing all the pro- spective properties of the future organism. For that reason, their sub- microscopic structure is of particular interest. The results attained in this direction are still rather scanty; but there are already some interesting electron microscopic investigations on the fine-structure of gametes which are reviewed below. a. Spermatozoa The tails of certain spermatozoa are positively birefringent, whereas their heads are negative (Schmidt, 1937a). The inference is that the anisotropy of the tail is due to protein fibrils, that of the head to the inclusion of orientated nucleic acid (Fig. 125a, p. 228). The head of the sperm being too thick for the transmission of electrons in the electron microscope, only details of its surface can be explored; but the thinner tail offers excellent conditions for such an investigation, and the submicroscopic structure of this part of spermatozoa is now thoroughly known. Using the microscopic information available, Bretschneider (1949b) has drawn the diagram of Fig. 137 as a result of his electron microscopic investigations. A strong nuclear membrane of protein fibrils causes the characteristic shape of the head, which contains the chromosomes. It is enveloped by a thin layer of cortical cytoplasm. The apex is covered by the so-called head cap consisting of a highly hydrated gel that plays an important role in fertilization (Bret- schneider, 1950b). Its distal end is marked by a sharp line in the cortical plasm. The basal part of the head is covered by a very thin sheath, the head tunica. There is a collar formed by a ring-shaped membrane around the base of the head, where the tail is fastened. The tail consists of 9 microfibrils into which the axon of the GAMETES 275 flagellum can be split. Eight of these microfibrils are arranged in a tube and their ends are connected to the base of the head. They surround the ninth microfibril. This central fibril is fastened to the centrosome which is situated in the crater-shaped base of the head. These 9 microfibrils are enveloped by a helical sheath consisting of a double spiral, each band of which is about 1 70 m// thick. The spiral body origin- ates from mitochondria; it is rich in lipids. It ends with the so-called ring of Jensen who had discovered the spiral body in the ordinary micros- cope (1887). Further on the axon is covered by a thin cortical membrane, which again has a helical texture (tail spirals). It consists of microfibrils about 50 m// thick with a low pitch making about 150 spiral windings around the axon. The terminal part of the tail protrudes from the cortical membrane showing the uncovered axon. Usually this part is slightly curved or sharply bent at the end (Fig. 137). When bull sperm is dried, the microfibrils of the axon fall apart, forming a tiny brush which is an artefact. In human sperm this is not the case. It is remarkable that the number of 9 microfibrils is not only characteristic for the sperm tail of many vertebrates (e.g. Corregonus; Rotheli, Roth and Medem, 1950), but also for some invertebrates investigated, such as sea-urchins and coleoptera (Bretschneider 1948). In ram spermatozoa 12 microfibrils have been found (Randall and Fried- lander, 1950), 6 of which form a tube surrounding a sixfold central fibril. Minute details of the connec- tions of the tail fibrils to the head and the compli- cated helical textures of the spiral body and the corti- cal membrane are also described by these authors. In algae there are spermatozoa with hairy flagella (in German "Flimmergeisseln"). With Euglena and Monas the hairs have been discovered in the ordinary microscope (Fischer, 1894). Fig. 137. Fine- structure of the sperm (from Bret- schneider, 1949). I head cap ; 2 chro- mosomes ; 3 head tunica (external layer) ; 4 ring- shaped membrane ; 5 centrosome; 6 articular strands ; 7 axial filament; 8 double helix (Jensen's spiral body); 9 Jensen's ring; 10 cortical helix; 11 terminal piece. 276 FINE-STRUCTURE OF PROTOPLASM II They are of the same order of size as bacterial cilia and, like them, can only be made visible under the light microscope by methods which increase their width, for example, by the use of an apposition stain. They are also described for aquatic fungi, certain brown flagel- lates (Chrysophyceae) and the zoospores of the Heterocontae among algae. In the electron microscope these hairs are very conspicuous (Brown, 1945; Foster and co-workers, 1947; Houwink, 195 i). Manton and Clarke (1950) have discovered that the longer one of the two flagella of the spermatozoa of Fucus is also hairy. It will be an interesting task to find out whether these hairs are active, like the bacterial cilia, or whether they are passive microfibrils split off the fibrous flagellum in order to increase its propulsive power. b. Eggs There is a wealth of information on the birefringence of the cortex (Fig. 138) of the sea-urchin egg (Monroy, 1945; Monroy and Mon- talenti, 1947), indicating that this layer is a lipo-protein system Fig. 138. Tubifex egg (from Lehmann, 1947). (Ohman, 1945). After fertilization, the double refraction disappears for 15-20 minutes, indicating an activation of the cortical layer by temporary hydration and disorientation. Similar structural changes 5 GAMETES 277 have been observed in the dark iield microscope (Runnstrom, 1928/29). The contents of the egg are liquid. They can be stratified by cen- trifugation into layers of yolk, fibrillar cytoplasm and enchylema with mitochondria (Fig. 113, p. 195). The cytoplasmic fibrils are double refracting; they carry the ribonucleic chromidia (Monne, 1946a). The egg of Tubifex has a much thinner cortical layer, which is easily destroyed by lipid solvents. The ground cytoplasm consists of fibrils which are beaded by chromidia. In the electron microscope the chromidia measure 0.15 /^ (Lehmann and Biss, 1949). The fibrils, whose diameter is smaller than o. i fx, form a coarse meshwork which harbours the microscopic yolk granules (ca, 2 fi diameter). This gel can loosen its junctions, so that the fibrils display protoplasmic flow, which is the case during anaphase and telophase of mitosis. In this state the egg content is liquid and can be stratified by centrifugation in the same way as described for the sea-urchin egg. At the two poles the protoplasm of the Tubifex egg is clearly differentiated into regions of animal and of vegetal cytoplasm. These differentiations are microscopically visible because they contain strongly basophilic granules ^Fig- 158)- Prior to fertilization the cyto- plasm of either pole can be forced across the cell by centrifugation and united with the cytoplasmic opposite pole (Lehmann, 1948). There is no mixture with the central fibrillar cytoplasm. The stratification produced is stable in Tubifex eggs, whereas in other cases, as e.g. in Limnaea eggs {Molhisca), the original arrangement is restored by protoplasmic flow (Raven and Bretschneider, 1942). Tubifex eggs with displaced polar cytoplasm can develop normal embryos. But they do not do so when the centrifugation has been applied too early. Lehmann (1948) thinks that the polar regions differentiate at the expense of the yolk and that their development is interrupted after the centrifugal displacement. Since, during the cleavage of the egg, the polar cytoplasm can be traced into definite somatic cells, it can be shown that in germs with abnormal develop- ment those somatic cells contain too small a portion of polar cyto- plasm. This shows how local regions of the egg are capable of inducing the development of definite parts of the germ. For that reason, there are not only multicellular organizers which control the organo- genesis during the development of the embryo, but there are already regulating systems on a lower scale inside the egg cell. In this way 278 FINE-STRUCTURE OF PROTOPLASM II it is shown that the morphologically differentiated parts of the cyto- plasm fulfil different physiological tasks. Lehmann (1948) distinguishes the following systems in the cyto- plasm of the egg of Tubifex which, during the development of the germ, play a definite role of their own: the cortical layer, the animal pole cytoplasm, the vegetal pole cytoplasm, the fibrillar central cyto- plasm and the cytoplasm round the nucleus (Fig. 138). It would be of great interest to know the submicroscopic structure of these different types of cytoplasm. Lehmann (1950) has started this important electron microscope investigation with the following pre- liminary results. The fibrillar cytoplasm consists of coarse beaded fibrils carrying the chromidia and enclosing the yolk granules as mendoned above. The polar cytoplasm has quite a different character; it is a dense mass of globular elements of 30-100 m^t diameter. These globules can associate and form a gel. As the polar cytoplasm of the egg is later transferred to ectodermal and mesodermal cells, they have been investigated individually. The ectodermal cells contain similar globules (30-100 m^<), but the mesodermal cells produce large ellipsoidal globules of the dimensions 600 x 300 m// or 300 X 200 m^. It is open to discussion how these large particles evolve from the smaller globules in the polar cytoplasm. From their density in the electron micrograph they are thought to contain phosphorus. This together with their microscopical size (ca. 0.3 fj) makes it look as though they are related to the basophilic granules which characterize the polar cytoplasm in the ordinary microscope (Fig. 138, p. 276). It is strange that they should not have appeared on the electron micro- graph of the polar cytoplasm. All these large basophilic granules with a high phosphorus content are probably not structural elements at all, but the seat of important metabolic processes. III. FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES The distinctive feature in the structure of living protoplasm is the absence of homogeneous lattice regions, whereas the intrinsic struc- ture of protoplasmic derivatives is as a rule conditioned by the ar- rangement of the molecular elementary units in some lattice order. o This is due to the fact that protoplasm is made up of many varying kinds of molecules (including specifically different polypeptides), whereas the high-polymer constituent of a protoplasmic derivative generally consists of one particular kind of macromolecules which combine to form an orderly pattern with comparative ease and thus lends itself to X-ray analysis. For this reason we are much better informed on the submicroscopic structure of these mesoplasmatic, metaplasmatic and alloplasmatic cell constituents than on the intrinsic structure of the living substance. Thus, while we can only trace the intrinsic structure of protoplasm in general outline, we have abundant quantitative data concerning the ultrastructure of highly diiferentiated cytoplasm, frame and reserve substances. This part of the monograph will deal with those structures disclosed up to date and will be con- cerned less with the physico-chemical than with the biological questions inherent in the theory of microstructure. The macro- molecular substances making up the bulk of the structures concerned are mentioned within brackets after the sub-titles. § I. Carbohydrates, Chitin and Cutin a. Meristen/atic Plant Cell Walls (Cellulose) The primary cell wall. There is a physiological and histochemical difference between the primary cell wall of vegetable meristems and the secondary membranes of fully grown tissues. It is mainly in their surface growth that this difference stands out. the secondary wall layers being, on the contrary, deposited by apposition against the ex- panded primary wall during the corresponding growth of the mem- brane in thickness. In many respects, therefore, the primary mem- 28o FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III branes behave very differently from the strong secondary walls. They lack microscopic lamination and iibrilization. Since they represent the intermediate membrane between neighbouring cells, they consist of three lamellae, viz., the original middle lamella produced from the cell plate during cell division, and the two primary walls added on to it. Another important point about meristematic cell walls is that no cellulose can be identified microchemically in them (Tupper-Carey and Priestley, 1923). Gundermann, Wergin, and Hess (1937) nevertheless detected by X-rays the fibre period of cellulose in the elongating cells oi Avena coleoptiles (after removal of the epidermis). As their photographs show only the interferences of the lattice planes perpendicular to the chain axis, evidently the cellulose strands present are either poorly crystallized or the X-ray pattern is disturbed by the large amount of pectic, hemicellulosic and other non-cellulosic wall substances. Thimann and Bonner (1933) found by analysis 42% of cellulose in dried A.vena coleoptiles but, just as in Heyn's X-ray investigations (1933, 1934), this percentage includes the epidermis with thickened walls (Fig. 140b, e, p. 284). Although unthickened meristem walls contain less cellulose, they certainly contain an already cohesive, fine framework of cellulose strands. Seeing that the cellulose is masked by other constituents of the membrane (see p. 287), it is particularly fortunate that its presence can be betrayed by its bire- fringence. Pectins, which accompany cellulose, have only very rarely been found to show birefringence in plants (Roelofsen and Kreger^ 1951). The view I advanced (1935 b) at the International Botanical Con- gress held at Amsterdam, to the effect that quite young meristematic cell walls already contain a submicroscopic cellulose framework, was at first disputed by Hess and his co-workers, though they overlooked the birefringence of these cell walls (Hess, Trogus and Wergin, 1936). Later, however, they admitted that cellulose can be identified by X-ray after cold water extraction, since, after the removal of water- soluble intermicellar substances, collective crystallization of ex- ceedingly thin strands of cellulose takes place (Hess, Kiessig, Wergin and Engel, 1939). Birefringence enables the investigator to detect when, during the formation of the young membrane after the division of the cellj cellulose first makes its appearance. Becker states (1934) that the so- I CARBOHYDRATES, CHITIN AND CUTIN 281 called cell plate in the phragmoplast of the staminal hairs of Trades- cantia first becomes visible as droplets exhibiting a Brownian move- ment. They do not, he says, move along the spindle filaments, as is assumed by others, but are formed, just where they are, by dissoci- ation from the dense plasm (Becker, 1935). The drops adhere laterally and form a grained isotropic membrane which, however, does not at first touch the side walls and shows a pectic reaction (coloration with ruthenium red). Plasmolysis reveals its independence. From the moment when this system has grown completely through the phrag- moplast and has reached the wall of the mother cell, this diaphragm becomes visible between crossed nicols. Apparently the phragmoplast, split into two halves, immediately generates cellulose on its surface where it is in contact with the new membrane. It seems to me im- probable that a cellulose frame would develop from the droplets described by Becker. It is also difficult to understand how proto- pectin could be formed from liquid drops. I therefore suspect that the drops are water of hydration liberated when high-polymeric chain molecules are built up in the cell plate from sugars of low molecular weight. The fact that the microvacuoles are dyed vitally with basic dyes (neutral red) does not invalidate this view, since they may quite conceivably contain water-soluble components, though they can scarcely harbour insoluble high-polymeric material such as proto- pectin or cellulose. These wall substances must be formed submicro- scopically in the phragmoplast and do not become visible until a microscopic system of protopectin has been built up, against which cellulose mixed up with protopectin is then immediately deposited on both sides. Hence the original middle lamella and both primary walls are already present in this very young state, but presumably all three membranes increase in thickness before surface growth begins. Cell elongation. The submicroscopic morphology of elongating cell walls is familiar. All meristematic cells capable of elongation are of tubular texture, as has been demonstrated in the case of Avena cole- optiles (SoDiNG, 1934; Bonner, 1935), of the staminal filaments (ScHOCH-BoDMER, 1936; Frey-Wyssling, 1936c), thc rapidly growing sporogonous stem of the moss Pellia (Overbeck, 1934; Van Iterson, 1935), to mention only a few. Likewise cotton hairs (Wergin, 1937), bast fibres and all derivatives of cambium (Meeuse, 1938, i 941) possess extremely thin, scarcely visible primary walls of tubular texture. The 282 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III cellulose framework of a wall of this kind is illustrated in diagram by Fig. 139, as derived from the birefringence and iodine dichroism o^ Euphorbia latex tubes (Frey-Wyssling, 1942). It should be borne in mind that with tubular texture the cell wall is negative with respect to the cell axis. When elongating tissues are stretched by mechanical means, the birefringence of their cells changes and becomes positive (Bonner, 1935); but if they extend through growth they remain nega- tively birefringent. We have to ask ourselves, therefore, why the cell wall optics of artificial and natural extension should be opposite. The electron microscopy of primary cell walls has disclosed a texture of cellulose strands al- most identical with the diagram of Fig. 139 drawn on the basis of indirect methods (Frey- Wyssling, Muhlethaler and Wyckoff, 1948; MiJHLETHALER, 1950a). The Strands are the same as the microfibrils observed in secondary cell walls (p. 105); their diameters are almost iden- tical. This had not been expected, because the fraction of cellulose is only a very small portion of the total amount of wall substances in primary walls (Table XXV, p. 287). The microfi- brils form systems which cross at different angles, but mostly so that an angle smaller than 90° points in the transverse direction of the cell. This causes the optical negative reaction in the polarizing microscope. A new fact, however, was also found, viz., that the microfibrils are not stratified in superposed planes but are interwoven, just as in a textile fabric. This is the reason why primary walls do not show any lami- nation and cannot be broken down into fibrils. On the other hand, the question arises as to how such a woven texture can grow in area. An investigation into the surface growth of these membranes has therefore been started. Plant cytology distinguishes two different types of cell elongation, termed tip growth and cell extension. Tip growth is considered to consist in the addition of new areas to the existing wall at the distal cell end, such as in elongating root hairs, cotton hairs, pollen tubes, fungal ^ k absorption coefficient, n refrative index. Fig. 139. Tubular texture of latex tubes (from Frey-Wyssling, 1942)^. I CARBOHYDRATES, CHITIN AND CUTIN 283 hyphae etc. On the other hand, very rapidly expanding cells in the tissues of coleoptiles, hypocotyls, radicles, staminal jfilaments etc. were thought to elongate by increasing their cell surface along its total length owing to passive extension accompanied by active intus- susception. The process of the addition of new microfibrils to the existing texture in tip growth is difficult to observe in the electron microscope. In growing root hairs the apex appears to be covered by a felt of cellulose microfibrils which stiffen the slime around these cells (Frey- Wyssling and Muhlethaler, 1949 b), and those of the pollen tubes (VoGEL, 1950) or of sprouting sporangiophores (Frey-Wyssling and Muhlethaler, 1950) are so intensely cutinized that the cellulose texture is obscured. Cells which grow in water do not present these difficulties. In the end cell of a Sp'iro^'ra thread the microfibrils are not intermeshed (Fig. 86b, p. 128). The tip consists of loose longi- tudinal microfibrils which represent a kind of warp. At their distal end these microfibrils seem to be free, whilst at their base they are tied together by transverse microfibrils which function as a weft. In this way a woven texture results. Soon the number of transverse microfibrils exceeds that of longitudinal fibrils, thus producing the optical negative reaction of the fully grown primary wall. In order to investigate the so-called extension growth, elongating coleoptiles were macerated and the isolated cells duly prepared and observed in the polarizing and the electron microscope (Muhle- thaler, 1950b). The result of this research is very surprising. It transpired that there is no extension of the wall in its total length, but the cell elongation is due to a rapid bipolar tip growth. This is illustrated by Fig. 140. Picture a) shows an expanding parenchyma cell of the oat coleoptile stained with benzoazurin in the polarizing microscope. The dichroism of this dyestuff produces deep coloration when the direction of the bulk of the microfibrils coincides with the vibration plane of the polarizer. It is seen from Fig. 140a that there is a heavily stained cell body with pits from which a long thin-walled outgrowth protrudes. In the cell body longitudinal ribs are visible which correspond to the cell edges. Fig. 140c gives a detail of such a rib with the adjacent pitted primary wall in the electron microscope. It is evident that a wall fortified by numerous parallel textured ribs cannot be extended in the longitudinal direction. Therefore, an ex- 284 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III Fig. 140. Elongating cells iii Arena coleoptilcs {c-e electron micrographs), a) Elongating parenchyma cell, 200:1; b) elongating epidermal cell, 630:1; c) face and edge of a parenchyma cell, 8000: i; d) tip of an elongating parenchyma cell, 8000: i ; e) tip of an elongating epidermal cell, 8000: i (from Muhlethaler, 1950b). I CARBOHYDRATES, CHITIN AND CUTIN 285 tension growth in the classical sense of such a cell is not possible. Growth in area is only realized in the two polar outgrowths, of which only one is visible in Fig. 140a. The tip of such a process seen in the electron microscope is shown in Fig. i4od. It is open, and evi- dently the same weaving of a transverse weft into a longitudinal warp takes place as was described above. RoELOFSEN (1951b) finds an axial orientation of the microfibrils on the outer surface of the primary wall of cotton hairs and a tangential orientation on the inner surface. He thinks that the outer fibrils have been oriented by cell extension. It is more likely, however, that these longitudinal microfibrils represent the "warp" as seen in Fig. 86b (p. 128). The impossibility of wall extension is even better illustrated by the epidermal cells (Figs. 140b, e). They elongate in the oat coleoptile about 150 times, (Frey-Wyssling, 1945a), but during the whole time of this rapid growth, which lasts four days, there is the compact outer wall, several // thick, characteristic of the epidermal cells of plants. The electron microscope discloses tip growth, not only for the thin- walled interior part of the epidermal cell, but also for that very thick exterior wall. It is an amazing thing that, simultaneously, in one and the same cell, a tubular texture should be laid down for the interior faces of the cell w-all and a parallel one for the thick exterior faces. This fact argues against any simple physico-chemical origin of cell wall textures comparable to that of molecular surface films. There are unknown morphogenetical principles inherent in the cytoplasm building the wall. From Figs. i4od and e it w^ould seem that the cytoplasm oozes out of the cell in order to weave its wall, not only from inside, but also from outside. The discovery of bipolar growth raises the question whether there is any intercalation of microfibrils by intussusception. Hitherto the growth in area was considered to consist in local expansions of the wall and concomitant insertion of new cell wall substances into the loosened area. The bipolar growth does not favour such a view, because it consists essentially of an addition of a new area to the existing wall and not in a general enlarging of the cell faces by internal growth. However, there are growth phenomena, such as the enlarge- ment of the cross-section of plant cells, which cannot occur by the simple additon of new wall areas. This growth in area consists in -86 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III pushing the cellulosic microfibrils apart by local plasmatic growth (mosaic growth; Frey-Wyssling and Stecher, 195 i; Bosshard, 1952). Actually the insertion of additional cellulose microfibrils into the existing fabric is not quite as difficult as it seems from the electron 0.03(1 Fig. 141. Cellulose frame in living cell walls (from Frey-Wyssling, 195 i). micrographs. The cellulose texture observed represents only 2.5 % by weight of the growing cell wall; in the living state it contains 92.5 % of water of hydration and only 7. 5 % of wall substances, of which 2/5 are pectins and hemicelluloses which are removed when the cells are prepared for examination in the electron microscope. On the basis of these figures and the known diameter of the cellulose microfibrils the diagram of Fig. 141 has been drawn (Frey-Wyssling, 195 i), which shows how much space is available for living cytoplasm [Christian- sen and Thimann (1950) find 12 ^% protein in the primary wall of pea seedlings] and highly hydrated accompanying substances in a CARBOHYDRATES, CHITIN AND CUTIN 287 primary cell wall. It also rules out the possibility of direct interference by auxin with the cellulose frame. The increased plasticity of elongating tissues (Heyn, 193 1 ; Soding, 193 1 ; ZoLLiKOFER, 1955) is probably due to the bipolar protrusions of the cells, and the effect of different ions on the cell elongation (Wuhr- MANN, 1937) must be sought in the influence on the cellulose-synthe- sizing cytoplasm. TABLE XXV CHEMICAL COMPOSITION OF MAIZE COLEOPTILES IN mg/COLEOPTILE (blank AND FREY-WYSSLING, I941; WIRTH, I946) Length of coleoptile in mm 9 52 55 55/6 Lipids Sugar Hemicelluloses . . Cellulose Pectin Protein Ash 0.040 1. 016 0.251 0.191 0.052 0.5 10 0.160 0.701 2.65 1 0.975 0.950 0.272 1.018 0.500 0.975 5.704 X.571 1.616 0.580 1. 651 0.444 0.162 0.951 0.228 0.269 0.095 0.272 0.078 Sum Total dry weight . 2.200 6.845 2.545 6.755 1 12.521 12.400 1 2.055 2.067 Forces of growth. The classical cytologists considered the turgor pressure to be the driving force of cell elongation in plants. The cell expansion was ascribed to water absorption only. This is not the case, however, as seen from Table XXV, where the chemical composition of expanding maize coleoptiles is summarized. Since there are no cell divisions when the coleoptiles elongate from 9 to 55 mm, each cell must increase all its constituents, in the same proportion as indicated by Table XXV. The increment of cell substances appears to be very considerable (Blank and Frey-Wyssling, 1941, 1944), being almost proportional to the cell elongation. If the figures relating to the 5 5 mm coleoptile are divided by 6, the values for a coleoptile section of 9 mm length are obtained (Table XXV, last column), which compare 288 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III c o 50 - 40 - 30 - 20 - W ■c c J' 400 300 200 o S o e CO o 40 30 20 favourably with those of the 9 mm coleoptile; a real increase per mm occurs for lipids, pectins and above all for cellulose, which is compen- sated by a loss in proteins and ash. This investigation proves that cell elongation is accompanied by a most intense metabolism^. Osmotic phenomena are only accessory manifestations of that metabolism; they are never the cause of any growth. BuRSTROM (1942) has carefully studied the osmotic conditions during cell elongation in wheat QOnjAt root. It is seen from Fig. 142 that the turgor pressure temporarily decreases during the lengthening of the cell. To raise it to its initial level, osmotic material has to be brought into the cell. Since ener- gy is required to transport ma- terial (Arisz, 1943), there must be considerable respiration dur- ing the elongation of the cell (Bonner, 1936b). This proceeds, therefore, not only by means of osmotically accumulated poten- tial energy, but chemical respira- tory energy is needed as well. Turgor extension is at its greatest at the moment when turgor pressure is at its lowest, from which it follows that the wall then has its maximum elasticity (Frey-Wyssling, 1948 a, b). Af- terwards elasticity is obviously reduced by the stiffening of the new wall areas (Fig. 140a, p. 284). It is curious that, despite the turgor, the stretchable bipolar cell outgrowths show no tendency to become spherical during the ex- tension. This is due to the submicroscopic tubular texture of the cells, 1 BuRSTROM (195 1) produces evidence showing that cell elongation and increase of dry matter are nevertheless physiologically separated processes. m -10 0 10 20 Time units Fig. 142. Osmotic conditions during the elongation growth of single cells in wheat root (compounded from various illustrations in BuRSTROM, 1942). Abscissa: Time (time unit is duration of mitosis in the tip of the root). Ordinates: a) Length of cell in^; b) turgor extension in yu ; c) turgor pressure in at.; d) osmotic material per cell in (lo//)^ times at. (From Frey-Wyssling, 1945 a). CARBOHYDRATES, CHITIN AND CUTIN 289 which resists any such tendency. The microfibrils of the cellulose frame, which encircle the cell horizontally to obliquely, have con- siderable tensile strength which is comparable to that of bast fibres and is due to primary valency bonds. In the axial direction, however, these fibrils are held together only by interfibrillar substances of much -2r- fa tit.* itit' Pa O) b) Fig. 143. Wall tension in cylindrical cells, a) Anisotropy of the strength F and of the wall tension p axially (index a) and tangentially (index t) ; b) derivation of longitudinal (Pa) and lateral stress (pj). 1 length, r radius of the cell, d thickness of the cell wall. weaker solidity. Consequently, a cylindrical cell of tubular texture has less strength axially than tangentially (Fig. 143^)- It is therefore not difficult to understand that the elastic extension by the turgor occurs preferentially in the axial direction. The turgor tension in the cell wall likewise differs according to the direction, and in the same sense as the strength of the wall. As the equation (Castle, 1937b) wall tension p X cross section of wall = turgor pressure T x liquid cross section applies, we have p3-(2 7rrd) = T-TiT- pr(2ld) =T-2rl for the axial (pj and tangential (p,) wall tension, where d is the wall thickness, r the radius and 1 the length of the cylindrical cells (Fig. 143b). The resultant ratio of p^ to p^ is 2:1, i.e., the tangential wall tension is double the axial wall tension. Although the lateral stress in the extending cell is twice the longitudinal stress, it grows in length only. This is possible if the F^: F^ strength ratio is above 2, as there is every reason to think it will be, since primary valence bonds are chiefly responsible for Ft, whereas cohesive forces, which are ten times smaller, determine F^^ (see Tables III, p. 31, and FV, p. 32). 290 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES HE This means to say that elastic cell extension, instead of giving wa^r to the lateral tension, follows the weaker longitudinal stress. Castle (1937b) thinks that the larger tangential stress favours the transverse orientation of the cellulose strand and thus causes the tubular texture. Careful examination of Fig. 140 (p. 284), however^ leads one to conclude that no such mechanistic process can explain the very complicated facts of the submicroscopic morphogenesis which is observed. CoRRENS, who noted the predominance of lateral stress in cylindri- cal cells as far back as 1893, came to the conclusion that "the existing stressing effects" in the micellar texture of laminated membranes of filiform algae "cannot be responsible for their orientation" (1893,. p. 284), since laterally and longitudinally orientated systems occur alternately. A further argument which proves the relative unimportance of the turgor pressure in growth problems is the study of energetics during cell elongation. Assuming the elongation to be really an elastic stretch, which is later fixed by intussusception, the work involved in wait extension can be calculated (Frey-Wyssling, 1948a, b). It turns out that this work is only i/iooo to i/ioo of the total energy produced b}~ the cell when the sugar content of its vacuolar sap is respited. For this reason, there must be other fundamental processes, such as- transport of substances and biosynthesis, which cause growth, and the problem of morphogenesis remains as enigmatic for submicroscopic morphologists as it was for microscopic cytologists. The secondary cell wall. According to Van Iterson (1927) the sub- microscopical texture of the secondary cell wall depends on the direction of flow of the protoplasm depositing the laminae of ap- position. Currents of protoplasm can, in fact, be observed to circulate,, depositing rings or bars during vascular formation. Van Iterson (1937) furthermore tries to explain the direction of flow causally. It is, he says, principally axial in the staminal hairs of Tradescantia, for example, since, owing to the tubular texture of the cellulose mem- brane, the cells tend to elongate. However, the outer cuticular layer with fibrous texture impedes extension, but there is pronounced elon- gation the moment the cuticularized outer layer of withered flowers bursts. On the basis of these observations it was inferred that, owing to the tubular texture of the primary wall of embryonic fibres, the CARBOHYDRATES, CHITIN AND CUTIN 2QI protoplasm likewise circulates in an axial direction and the nascence of the fibrous texture of the secondary wall could be explained as being causally mechanistic. Van Iterson now goes so far as to suggest as an explanation for the crosswise layers of the Valonia cell wall (brought into prominence by X-ray investigation) that the proto- #c --©t 'S-- jjUI/> ''Tr- m 'i(,\<- ^^■- ;@c '!wf ^'©'4 ''<^ ^^\ll/ .-.Ml/, ''/!«'> N.11// ■J.^^ 5@? ^,©f '"/n^" '/;i^ .^5Ii!^ <,M'/^ ;@5 W <^.uHui^:i:>' .,^ ~-^i!S:^miiiiMHill>' ^__,,^-^ui,Mm,-;T^ •'^""'"'^^^rrrr;^ ''-^^i:::::;/^!,!.!;;;:^-- ^_.,,,,-;;;5i^MiMiS7^7;~^ •^<:::^mmiMi!^ -:l:^. "•-^^''IIIMIIj^ — ■ .„.--^SiTii"'"i""7ii^ -^iii''"'i7u7j^^ ■i^^'l/IIMIIIjl^ii^^^-'' .„-. "^ — .JJ^iimww^;^^ .a!!!;i"iiMiiiiii;;:!Ii>-' --ss^ ""^^^^5sJ^- '^^.aSg'"""!^^^^^^^ -"""^""""^^Tm)^ -ssr«"""i"/inij~.«.^ •'^^.miiii.nmiff:^-^ -<:;ii>^'^^''^^^^ ">. ■<::2imMMnmin;;2>^ ^^^iMiiM,rtiii;~~j^ ^) Fig. 144. Diagram of intercalation of wax in meristematic cell walls of tubular texture. a) Radial section; U) tangential and cross-sections. plasm is forced to change its direction of flow by about 90° after the deposition of every layer; for the tendency of the cell to expand is always perpendicular to the direction of the iibrillae of the newly formed layer, for which reason the flow of plasm is supposed to be passively directed cross-wise over the youngest lamella. With all due admiration for Van Iterson's reasoning, and con- ceding a certain contributory role to the forces he has discovered, it can scarcely be said that mechanistic theories of this kind are at present adequate to resolve the mysteries of morphogenesis. For there are several facts of observation which do not come within the compass of causation. For instance, out of similar cells near the cambium, to the primary wall of which tubular texture is ascribed, are differentiated on the one hand fibres with spiral texture, which may have been generated as suggested above, but on the other hand, vessels with tubular texture. This might be due to the fact that the fibres have tip growth, whereas the vascular members have not; it would then, 292 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III however, have to be explained why in one case the extensibility of the primary wall of tubular texture is utilized, while the cylindrical cells of the vascular members follow the unconventional course of growing in girth instead of in length. It is difficult to avoid assuming the existence of internal formative forces when the purposiveness inherent in each individual cell development becomes apparent again and again. Intercalation oj wax. The discovery by X-ray of the intercalation of wax has enriched our knowledge of the structure of the primary cell walls. In young cotton hairs, A.vena coleoptiles without epidermis and many meristematic tissues, Hess and co-workers (1936) found X-ray interferences corresponding to periods of 60 and 83 A. By extraction these substances were isolated and identified as vegetable waxes (GuNDERMANN, Wergin, and Hess, 1937). They are comparatively short chains of the type QHgn+jCO-O-C^^Hoj^ + j, n and m amounting to about 24 or 32, as established for other vegetable waxes by Chib- NALL, Piper, and co-workers (1934). As these waxes produce far clearer interferences than cellulose, of which often only the fibre period appears, they must be assumed to be better crystallized than the cellulose chain molecules. The possi- bility therefore exists that waxes of this kind are in part the source of the birefringence of the primary cell walls. Pursuing this problem as presented by the meristematic cell walls of A.vena coleoptile, K. and M. Wuhrmann-Meyer (1939) established that the birefringence is affected by the fatty wax component susceptible of extraction by pyridine. Though this effect is, admittedly, lacking in the radial sections through the cells, it appears in the tangential and cross- sections. From this it may be inferred that the rod-shaped wax mole- cules are orientated at right angles to the microfibrils of the tubular texture; then there is isotropy on the radial section, whereas on the tangential and cross-sections we have a birefringence which is the reverse in character of that of cellulose, as will be clear from Fig. 144. The waxes being extremely hydrophobic and the cellulose chains very hydrophilic, there can be no direct contact between these two cell wall substances, so that an intermediate, polar substance is in- terposed (Frey-Wyssling, 1 93 yd). Possible molecules with hydro- philic and hydrophobic end groups are phosphatides (Hansteen- Cranner, 1926). Seeing that Thimann and Bonner (1933) found no I CARBOHYDRATES, CHITIN AND CUTIN 295 phosphatides in the membranes of Avena coleoptile, the question arises as to whether the wax alcohols and fatty acids in the primary walls occur in the unesterihed state, in which case their hydrophilic pole would be connected with the cellulose threads. It will be evident from Fig. 144 why the primary cell walls can be stained with fatty acid dyes, whereas the individual cellulose strands seem to be "masked". Physiologically this intercalation of wax results in the impaired permeability ot the wall to water, ions and lipophobic molecules, as these substances are admitted, not through the entire meshes of the intermicellar spaces but only through the hydrophilic regions in the vicinity of the cellulose strands. b. Cutini':(ed Cell Walls (Cut in) Mkrochemistry and optics of CMtini':(ed epidermises. The morphology of the thick cuticular layers of the leaf epidermises of xerophytes (Fritz, 1935, 1937) is particularly interesting, in that, although optically often appearing to be homogeneous, they contain at least four different membranous substances, the submicroscopic arrangement of which is known. Our starting point will be the optics, investigated by Am- BRONN (1888), of the cuticular layers which, in the polarizing micro- scope, behave in a reverse sense to the cellulose layers lying beneath them. The cellulose component appears optically positive with reference to the tangential direction of the cell wall, while on the contrary the cuticular layer is optically negative (Fig. 145a). Extern- ally, the epidermis is bounded by the almost isotropic cuticle and between the cellulose and cuticular layers is interposed a fairly wide isotropic layer of pectins (Anderson, 1928). Ambronn had already suspected that the optically negative reaction of the cuticular layers was caused by intercalated waxes, but this property was later attributed to the cutin. Madeleine Meyer (1938), however, demonstrated by careful micromelting tests (Fig. 145 b) that the negative birefringence derives from a fusible wax, while the residual cutin proves to be almost isotropic. In many cases, of which Gasteria is an example, a slightly positive birefringence, due to cellulose, makes its appearance after the waxes have melted out. Hence, besides the cutin, the cuticular layer must also contain cellulose and even pectins, which can be identified by ruthenium red. The optics of the longitudinal section discloses the fact that these four cell wall substances (Table XXVI) are not 294 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III TABLE XXVI CELL WALL SUBSTANCES OF THE CUTICULAR LAYERS Optical behaviour TTV referred to tan- Coloured by Solubility Disintegration Absorption gential direction by Cutin Isotropic Basic lipid dyes Insoluble NaOH saponification Strong Cutih waxes Opt. negative Lipid dyes Pyridine Melting above 220° C Lacking Cellulose Opt. positive Iodine-zinc chloride sol. (dichroism) SCHWEIZER reagent Hydrolysis Lacking Pectins Isotropic Ruthenium red Picric acid followed by H2O2 Hydrolysis Lacking Clivia Gasieria Yucca Dasylirion A 20^ 12° 4 (\ 12 3 4 [\ '-^yr* 12 3 4 2(f 40° Fig. 145. Cuticular layers of vegetable epidermises (from M.Meyer, 1938). a) Optics of longitudinal section. Ordinate: Relative strength of bire- fringence. Abscissa: i cellulose layer, pos. bire- fringent; 2 pectin layer, isotropic; 3 cutinized ''•' wall, neg. birefringent; 4 cuticle, isotropic, h) Hysteresis-melting curve of the cutin wax of Clivia, measured by reduction in birefringence of the section. Ordinate: Retardation in degrees of the Senarmont compensator. Abscissa: Temperature T in °C. 60'' evenly distributed over the thickness of the cuticular layer. In Clivia, for instance, only an inner zone — which iodine-zinc chloride solution tints dark brown — clearly contains cellulose. The waxes are in greatest evidence in the middle of the layer, so that it is there that 1 CARBOHYDRATES, CHITIN AND CUTIN 295 the retardation is at its most negative (Frey, 1926b). The wax content diminishes outside and the cuticle contains no wax at all, consisting •of pure cutin (Fig. 145a). In hydrophytes the cutinization of the epidermis is confined to a thin, optically isotropic cuticle. It is probable that all cell walls that are in contact with air are superficially cutinized, since Elsa Hauser- MANN states (1944) that the cells of mesophylls, which serve to ventilate the leaf, are covered with a submicroscopical film of cutin. Molecular stridcture of lipophilic cell wall substances. To understand the submicroscopical arrangement of the four cell wall substances in cuticular layers it is necessary to know the morphology of their mole- cules. We shall therefore have to consider briefly the chemistry of the waxes and of the very imperfectly known cutin. Unlike the enormously long cellulose chain molecules and the very long pectin chains, the waxes are, as already mentioned, short rod molecules of less than 100 A length. In the simplest case they consist of higher aliphatic alcohols, the corresponding fatty acids and higher paraffins. According to Kreger (1949), there is no stoichiometrical relation between alcohols and acids. Therefore, the plant waxes are only partly esters, the rest being mixtures of higher alcohols, paraffins and fatty acids, with a predominance of the first two. The alcohols and fatty acids have even-numbered chains between C24 and C34 (Chibnall, Piper and their collaborators, 1934); for instance, myricyl alcohol CgoHgjOH or cerotic acid C25H51COOH. If esters occur they have the same overall formula as fatty acids C^HanOg. On the other hand, the paraf- fins have odd-numbered chains between C27 and C31 (Kreger, 1949); e.g., n-noneicosane CogHgo. As indicated in the last formula, the mole- cules of plant waxes are always unbranched chains. Besides the aliphatic waxes the cutinized, and especially the suberized, membranes contain the waxes cerin and friedelin, which have a substantially lower hydrogen content. The inference therefore is that they contain aromatic rings and thus approximate the sterols, which represent the cycHc alcohols. Luscher (1956) states that friedelin and cerin contain an alcoholic OH group which can be acetylated or otherwise esterified, while the second constituent O atom is masked, presumably as a cyclic ether bridge. Thus friedelin and cerin are alcohols, not esters. On the other hand, they may possibly be esterified with other molecules in the membrane. Other- '.^6 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III wise nothing is known of their constitution. On saponifying the waxes of pine needles, Bougault and Bourdier (1908) obtained co~ hydroxyfatty acids (such as hydroxylauric acid and hydroxypalmitic acid) instead of simple acids and alcohols. Molecules of this kind possess two reactive groups; thus they can together form esters and grow to high-polymeric chains, as shown in Table XXVII. Their discoverers call these waxes "estolids". Their degree of polymeri- zation cannot be very high, as they are still soluble and fusible. TABLE XXVII lipophilic cell wall substances Aliphatic Waxes: Wax Acids: CH3. (CH,)„. C-O^ (CH,)^. CH3 ' Palmitic acid C15H31COOH Stearic acid Ci^Ha^COOH 0 Oleic acid Ci,H33COOH Linoleic acid Ci,H3iCOOH Arachic acid Q,H3,COOH (Chibnall and Piper, 1934; Cerotic acid Q5H51COOH LiJscHER, 1936) Higher fatty acids up to Wax Alcohols: QjHejCOOH Cetyl alcohol QsHsaOH Octadecyl alcohol QSH37OH Cer}l alcohol QeHssOH Myricyl alcohol QoHsiOH \ Higher alcohols up to C34H^90H Cyclic Waxes: Molecular structure : (LiJscHER, 1936) Estolids: -O-CCHa) -C-0-(CH,) -C-O- i! " !l o o (Bougault and Bourdier, 1908J Cerin QqH 5oO. Friedelin Q«H 7 60, Hydroxyacids : Sabinic acid (hydroxylauric acid) OH. ^11^22 COOH funiperic acid (hydroxypalmitic acid) OH- ^isHso COOH Suberin, Cutin and Sporopollenin: Molecular structure : Spatial network through ester and ether bridges (Zetzsche, 1932; LiJscHER, 1936) Suberin Cutin Sporopollenin Saponification becomes more and more difficult Decomposition Products of Sub er in: Suberic acid COOH- (CH,)6- COOH Phloionolic acid Ci,H3,,(OH)3-COOH Phloionic acid COOH- Ci6H3o(OH),- COOH Phellonicacid C2iH4.(OH).COOH Eicosancdicarboxylic acid COOH- (CH2)3„-COOH I CARBOHYDRATES, CHITIN AND CUTIN 297 The polymerization plan of the high-polymeric cell wall substances cutin and suberin must be similar to that of the estolids, since their hydrolytic and decomposition products ordinarily exhibit two or more reactive groups capable of esterifying or etherifying (dicarboxylic acids, hydroxycarboxylic acids, Table XXVII). This is the distin- guishing feature between the monomeric molecular residues of cutin and suberin, on the one hand, and the molecules of waxes on the other (LuscHER, 1936). Seeing that suberin is more readily decom- posed than the cutins (Zetzsche, 1932), it is probable that the degree of polymerization or of interlinking attained within it is lower than in the latter. It is presumably at its highest in sporopollenin, as this wall substance is exceedingly resistant to saponification and decay, so that the cell walls of fungus spores and grains of pollen are preserved for thousands of years in peat deposits. The isolated dicarboxylic acids (Table XXVII) may possibly be oxidized degradation products of higher hydroxyacids ; suberic acid, COOH- (CH2)6-COOH, for instance, results from the oxidative de- gradation of suberin. Probably not all the carboxyl groups of the carboxylic acids in the membrane are esterified, for cutin has some of the characteristics of an acid, or a high-polymeric anion (pro- nounced negative charge Brauner, 1930, selective cation perme- ability, staining by basic dyes). Since its behaviour is almost iso- tropic, it must be presumed that the linkage of the carboxyl and hydroxyl groups is not that of a linear chain scheme, but reticular in aU spatial directions as in lignin. Suhmicroscopic structure of the cuticular layers. It now remains to build up a picture of the mutual spatial relationship between the cell wall substances in the cuticular layers. A possible clue is afforded by the optical anisotropy of the suhmicroscopic particles of wax. If their form and optics were known, the orientation of the intercalated wax could be inferred from the nature of the wall birefringence. The wax molecules are rod-shaped and therefore, when spread on a slide, might be expected to be orientated and reveal something as to their intrinsic birefringence. Many waxes, like paraffin, fats, phos- phatides and other lipids, produce w^hat is known as a "negative streak" (Fig. 146 b), which might incline one to conclude that the wax molecules are optically negative with reference to their longi- tudinal axis. Such a conclusion is, however, inadmissible, since short- 298 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III redded molecules have a tendency to crystallize as thin platelets or lamellae (Fig. 146a) and, when spread out, these submicroscopic, often plastic, crystal lamellae are orientated. Longitudinally, they fall in with the direction of the stroke and the molecules then run perpen- dicular to the streak. In this wav the streaks of paraffin and beeswax ^^ a) + b) =1 c) W ^ =-c /". /\ /^ d) Fig. 146. Submicroscopic textures from optically positive lipid chains. P paraflfin mole- cules, E estolid molecules, W wax molecules, C cellulose chains, Cu interlinked cutin chains, a) Paraffin lamella optically positive; b) optically negative streak of paraffin; c) optically positive streak of pine-needle wax ; d) intercalation of cutin wax in the epidermis. are negative, but the molecules themselves are optically positive^. By analogy it might therefore be supposed that the molecules of the vegetable waxes which yield a negative streak are positive; but there are some waxes with a positive streak, as I found with estolids from pine needles (Fig. 146c). The streak test, therefore, tells us nothing definite and another method has to be resorted to, which consists in dissolving the waxes, in order that their molecules may be rendered independent of each other, and then testing their intrinsic bire- fringence in a flow gradient. Ambronn and Frey pointed out in "Polarisationsmikroskop" (1926, p. 167) that the only certain way of establishing the intrinsic birefringence of disperse particles is by using a rotary drum in ac- cordance with Kundt's system. Signer (1930, 1933) built a flow- ^ On p. 92 it is explained that double refraction cannot be attributed to a single molecule. So if we speak here of optically positive molecules, this means that the sign of the double refraction of a large number of molecules, made parallel by flow or crystal- lization, is positive. I CARBOHYDRATES, CHITIN AND CUTIN 299 birefringence apparatus of the greatest precision, in which, in spite of their Brownian movement, comparatively short rod-molecules can be orientated. It was with the aid of this apparatus that Weber (1942) determined the optical nature of wax molecules. The experimental evidence points to optically positive rod-molecules. Thus the molecules of the membrane waxes, like those of paraffin, fats and other lipids, are optically positive rodlets. Since the waxes, referred to the tangents of the cuticular layers, produce negative birefringence, their molecules must stand perpen- dicular to the surface of the membrane. So perfect is the orientation of the rod-molecules, that the outside layer of the epidermis of Clivia, seen from above after the removal of the cellulose layer underneath it, appears optically isotropic. Hence the cuticular layer possesses a radial optical axis. , After extraction of the wax, form birefringence is exhibited (form birefringence curves in M. Meyer, 1938), this, referred to the optical axis of the cuticular layer, being negative. This means that we have to do with lamellar birefringence; hence the wall layer consists of submicroscopic lamellae, in the texture of which, judging by all previous experience, the cellulose of the cutin layer must be involved. The optical analysis therefore suggests the presence of submicroscopic cellulose lamellae with exceedingly thin platelets of wax interposed, the wax molecules being orientated perpendicular to the cellulose chains (see Fig. 146 d). Now, in the presence of the water, present not only in cellulose, but also in cutinized cell walls, the hydrophobic wax molecules cannot come into contact with the hydrophilic cellulose chains. Thus there must be some intermediate polar substance, and that is the cutin. This wall material contains both hydrophilic (-OH, -COOH) and hydro- phobic (-CH3) groups and it may be assumed that the former incline more towards the cellulose, whereas the latter tend more towards the wax. We then have a scheme such as that represented in Fig. i46d. It can be seen in this model how the cell wall substances in the cuticular layers are placed one relatively to another: hydrophilic lamellae consisting of cellulose and probably also of pectins, layers of wax molecules in radial arrangement and, in between them, amorphous cutin in random orientation. Apart from the interposition of the wax, the morphological conditions are similar to those in lignification, where 500 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III amorphous lignin is intercalated between cellulose rodlets or lamellae. In both cases the cellulose is masked by the incrustation. For example, it is only with difficulty that the cellulose can be dissolved out of wood with Schweizer's reagent, and hitherto could not be eliminated at all in this way from the cutin layers. It is easier to saponify the cutin, or the suberin (Karrer, Peyer and Zegar, 1923; M. Meyer, 1958) and to hberate the cellulose. The scheme shows the relative positions of the four cell wall substances, not their quantitative proportions, these being very variable. Small or larger amounts of the carbohydrate wall substances, cellulose and pectins can always be identified in the inner regions of the cuticular layer; they are, indeed, often quite prominent. Further out, it is the waxes which are in greater prominence, with marked and sometimes complete decUne of cellulose and pectins. The outer layers probably consist of cutin and wax only. This is noteworthy as compared with lignin deposition, since cutin can obviously occur as an independent wall substance, whereas lignin is always found in company with cellulose. Finally, there are no waxes in the isotropic cuticle (Priestley, 1943), which, therefore, comprises only a thin pellicle of almost amorphous cutin. It would be interesting to discover the still quite unknown history of the development of this complicated submicroscopic system originating in a region remote from the protoplasm. Martens (1934) states that the cuticle is secreted in the fluid state and then coagulates in the air. This may also safely be said to apply to the cuticular layers. The cutinic acids would then be dissolved in a low molecular state, migrate into the wall and there polymerize. It is less difficult to under- stand the deposition of the low-molecular waxes, though even in this case it is necessary to assume that there is some special solvent, or that unesterified wax acids and alcohols migrate. This process is similar in nature to the excretion of waxes through the epidermis, where they form a granular, rod-shaped or scaly coating (Weber, 1942). Each component of the wall in the full-grown cuticular layer has its particular physiological function. By reason of its hydrophobic nature, the primary duty of the wax is to make these layers watertight. The cutin has a similar purpose, though in a less extreme degree, since its hydrophilic groups make it less hydrophobic and, therefore, it has I CARBOHYDRATES, CHITIN AND CUTIN 3OI a slio-ht tendency to swell. As the cutin layer strongly absorbs ultra- violet lio-ht (Frey, 1926 b) and retains this property even after the waxes have been extracted, it impedes any intensive ultraviolet ir- radiation of the mesophyll of xerophytes. As aliphatic compounds in general do not absorb ultraviolet light, there must be some unknown cyclic compound (cyclic waxes) in the cuticular layer. The hydrophilic quality of the lamellae of cellulose and pectins is responsible for the cuticular transpiration (Gaumann and Jaag, 1936) of the leaves, which occurs not only in hydrophytes, but also in xerophytic evergreens. The loss of water is a sign that the submicroscopic wax lamellae are not continuous, but that the hydrophilic (cellulose) and semi-hydro- philic (cutin) regions cohere and thus offer the water an outlet. c. The Chitin Frame (Cbitin) Chitin is a nitrogenous frame substance, primarily characteristic of the animal phylum of Arthropoda (Crustacea, insects). It also forms the membranous frame oi fungi (Harder, 1937; R. Frey, 1950). The behaviour of vegetable and animal chitin is identical, as has been proved for the sporangiophores of Phycomyces chemically, optically and by X-rays (Diehl and Van Iterson, 1935 ; Van Iterson, Meyer and LoTMAR, 1936). In the same way as the cellulose characteristic of autotrophic plants may be built by both bacteria {Bacterium xylinum) and by the animal class of the Tunicata (Fig. 86d, p. \z%), fungi are, inversely, able to synthesize an animal frame substance. One cannot go very far wrong by assuming that this similarity is connected with the heterotrophic life of fungi, which, like animals, have so much nitrogen to draw upon that some of it is deposited in the cell walls and is there immobilized. As there is, on the contrary, only a minimum of nitrogen in autotrophic plants, it cannot contribute to the form- ation of their frame substances; otherwise chitin, which is more resistant than cellulose in many respects, would certainly also occur elsewhere in the vegetable kingdom. Morphologically, the two frame substances are very similar in behaviour, as will be shown in what follows, the micellar frame of each being composed of very long chain molecules. Molecular structure of chitin. The structural unit of chitin is glucos- amine, i.e., a pyranose ring in which an OH group has been substi- tuted by an NHg group (Fig. 147a). It is not known whether the 302 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III position of this amino group at the z"'^ C atom corresponds to that of the OH group of the glucose or of the mannose ring (Itschner, 1935). An acetyl residue is linked with the NH2 group; thus, contrary to cellulose, there are here two side chains, viz., an OHCHo-group and a CHgCO-group. H NH^ H H OH /oh h\h oh /oh nh^ H\i_yOH ^\h_^/0H CH2OH CH2OH Glucose configuration Mannose configuration a) Glucosamine NHCOCH3 CH2OH NHCOCH3 CH2OH NHCOCH3 CH2OH NHCOCH3 CHpOH NHCOCH3 CH2OH h) Chain of chit in Fig. 147. Molecular structure of chitin. The acetylglucosamine molecules are linked glucosidically and form long chain molecules, each member of which is, according to Meyer and Mark (1930), twisted with respect to its preceding and succeeding neighbour by 180° (Fig. 147 b). X-ray photographs of the sinews of the spiny lobster and of the sporangiophores of Phycomyces show that the crystallographic elementary cell is rhombic, its di- mensions are 9.4 : 10.46 (10.26) : 19.25 A and it contains eight acetyl- glucosamine residues, viz., two to every four main valence chains, which traverse the crystal lattice (Meyer and Pankow, 1935). A different modification of chitin with the crystal lattice 9.32: 10.17: 22.15 has been found in Polychaeta and Mollusca. It has been termed i5-chitin, in contrast to the a-chitin of insects, Crustaceae and fungi (Lotmar and Picken, 1950). The fibre period 10.3 A is important, because it corresponds to the length of two pyranose rings and is identical to that of cellulose. This warrants the belief that the glucosan rings, hke the glucose residues of cellulose, are linked together by /5-glucosidic 1-4 bonds (see Fig. 147 b). Stihm'icroscopic texture of the chitin frame. Microscopically, the chitin sheath of the Arthropoda and the membranes oi fungi show lamellation and fibrillation, as is known to be the case in the cell walls of cellulose. By analogy, therefore, it may be assumed that fibrillation is realized CARBOHYDRATES, CHITIN AND CUTIN 303 in the submicroscopic region. The interfibrillar spaces in crustacea are filled partly with mineral substances, especially with calcium carbonate, while the membranes of fungi are encrusted with sub- stances rather of a carbohydrate or pectinous nature (which can be extracted by boiling for several hours with a ten per cent, solution of caustic potash). r; b?) e) Fig. 148. Types of submicroscopic texture with chitin as the frame substance, a) Crab sinew: fibrous texture, b) Interior of lobster shell: submicroscopic lamellar texture, with direction of fibrillae changing from lamella to lamella (i, 2, 3, 4, etc.); bj) cross-section, bg) plan, /r) Eggshell of Ascaris: foliate texture (Schmidt, 1936b). d) Conidiophores of Aspergillus: fibroid texture (Frey, 1927a). e) Conidiophores of Phycomyces: spiral texture (OoRT and Roelofsen, 1932). As with cellulose, the orientation of the rods of the frame is demonstrable by optical means, since the larger axis of the index ellipse of sections immersed in water or gh^cerol runs parallel to the submicroscopic chitin rodlets. This method reveals the same potential orientation as that actualized in cellulosic cell walls (Fig. 148). Chitinous tendons of crabs, lobsters, beetles, etc. are of an un- mistakable fibrous texture. Of all chitinous objects, therefore, they produce the most richly pointed X-ray diagrams and are thus the most informative as to the lattice structure of chitin. Optically, the fibrous texture is disclosed by the fact that the refractive power is considerably more pronounced parallel to the axis of the tendon than perpendicular to it, while something like isotropy prevails in the cross-sections of the tendon. This fibrous texture is to be inferred, not only from the birefringence, but also from the anisotropy of the absorption of light. Iodine-zinc chloride solution and Congo red stain decalcified and 304 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III cleaned chitinous tendons, as they do bast fibres^ dichroically; the direction of the stronger light absorption coincides, as in cellulose, with the fibre axis. The similarity in the dichroic coloration of chitin and cellulose is interesting in that it tends to show that the dichroism of these colour reactions characterizes not so much a certain chemical compound as its micellar structure with orientated inner surfaces. The egg-shell of ^j-^r^m provided Schmidt (1936) with an object in which the submicroscopic chitin rodlets scatter, thus forming a wall of foliate texture. The plan ot the eggs shows them to be isotropic, but the optical cross-section through the wall exhibits a negative spherite cross. This optical behaviour is produced by an arrangement of the submicroscopic ordered lattice regions as represented in Fig. 1 48c. The sporangiophores of Aspergillus niger must, from their optics, be presumed to have a fibroid texture with scattering (Fig. i48d; Frey, 1927a). We do not 3^et know, however, whether this membrane is stratiform like Phycowjces; for in that fungus, with particularly large sporangiophores several centimetres in length, Oort and Roelofsen (1932) found an outer primary skin of tubular texture, under which there is a thickened secondary wall layer of fibrous texture exhibiting slight scattering; it is by reason of its predominant bulk that only this appears on the X-ray photograph. It is assumed that at the core there is another, very thin layer of steep spiral texture (Fig. i48e). These results of the indirect methods are only partly corroborated by the electron microscope. The cell wall of the sporangiophore of Phycomyces consists of chitinous microfibrils which are similar to those in cellulose walls (Frey- Wyssling and Muhlethaler, 1950; Roelof- sen, 195 la). There is a homogeneous cuticle devoid of any structure, a primary wall with interwoven microfibrils and a thick parallel textured secondary wall (Fig. 86c, p. 128). Roelofsen differentiates the primary wall in an outer layer with a network texture and an inner layer with almost transverse oriented microfibrils. The texture of the uniform secondary wall runs almost parallel to the cell axis. There is no pronounced spiral texture and no special internal wall layer as had previously been found in the polarizing microscope (Fig. 1486). Spiral growth. The end of the sporangiophore is conspicuously of spiral growth (Oort, 193 i; Castle, 1937a, 1942). This fact can be verified by placing a mark above the zone of growth which was found ^ Walchli (1945). CARBOHYDRATES, CHITIN AND CUTIN 305 not only to travel upwards, but at the same time to rotate around the axis of the sporangiophore (Fig. 149). There is nothing in the submicroscopical texture of the primary wall which might account for this behaviour. Oort and Roelofsen (1932) state that the isolated wall is flabby and flexible and, as it tears impartially in all directions, is not ot parallel texture. This is con- firmed by the electron microscopic evidence. However, if the interior pressure in the zone of growth is artificially enhanced, the membrane bursts through a very steep spiral longitudinal tear, which may be at- tributed to the anisotropic states of tension in all tubular walls described on page 289. Artificial extension of the zone of growth is accompanied bv a rotation which, after relaxation, recovers. Thus the optics point to a woven tubular texture, while the mechanical properties require a spiral texture. Castle (1942) discovered ad- ditional complications; he was able to show that at first there is regularly a left tendency in growth, which then suddenly changes for an hour to a right-hand spiral and then reverts again to a left spiral. He tried to account for this by suggesting the preformation of both a left-hand and right-hand screw in the primary wall; that is to say, it would be a crossed system indistinguishable from the tubular texture. Preston (1948) has even developed a formula for calculation of the change of rotation from the elastic properties of the cell wall which alter during its diff'erentiation. But all these considerations are based on a spiral texture (Preston, 1934, 1936) which obviously is not realized in the growth zone of the Phjcomyces sporangiophore (Roelofsen, 1949/50, 1951a). Therefore, the simplest assumption is that intercalary growth in the zone of extension travels in a circle; this must be so, since the slender conical shape of the zone of growth could hardly be maintained if the surface grew simul- taneously on all sides. In fine with this is the fact observed by Oort and Roelofsen, viz., that in Phjcomjces Blakeskeanns var. piloboides Fig. 149. spiral growth of Phycomyces (from Castle, 1937 a). Zone of growth dotted; • marks to trace rotation. 3o6 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES HIT with a characteristically inflated sporangiophore, the sporangium does not rotate. It follows from this discussion, that there must be a local growth by^ intussusception, the mechanism of which is not yet fully understood. Crossed lamellar systems. The growth in area of membranes which; are obviously cross-textured is equally difficult to explain. An instructive example of a microscopically laminated structure is. provided by the inner layer of the lobster shell. As an entity, this, layer behaves like a uniaxial, optically negative composite body; i.e.,. seen trom the surface, it is isotropic. In cross-section, on the other hand, strongly birefringent (positive with reference to the lamellation) and isotropic layers are seen to alternate. Older investigators (Bieder- MANN, 1903) thought these lamellae possessed cross-wise fibrillation at right angles. Were this true, it should be possible to cut cross-^ sections at 45° to the two fibrillar directions through the composite body in which all the lamellae would show the same behaviour in the polarizing microscope. This, however, is not the case, for cross- sections, in whatever direction, through the lobster shell all invariably^ disclose the same pattern of lamellation. Schmidt (1924, p. 238) therefore assumes that the iibrillae in consecutive, very thin, parallel- fibred layers very gradually change direction, so that two layers at a certain distance from each other will contain fibrillae crossed at right angles, but those in between will contain fibrillae in any of the transitions from 0° to 90°. An arrangement such as this is indicated in Fig. i48bi (p. 303). This should be verifiable optically for, in the transition from lamella to lamella, the light retardation should drop- following a sine curve from the maximum value to nil. X-ray analysis, would likewise show whether all possible fibrillar directions are before us. It seems to me an important point that the hypothetical layers are submicroscopically thin for, were they of microscopical dimensions, it would mean that this is a comparable case to the spiral texture of cotton fibre ; that is to say, owing to the obliquely crossed layers, the top view of the interior layer of the shell could not be isotropic, but would have to transmit some light under all azimuths. Instead of assuming submicroscopic lamellae consisting of parallel microfibrils superimposed in different directions of orientation (Fig^ 148b, p. 303), it would be equally plausible to picture the micro- fibrils as interweaving. I CARBOHYDRATES, CHITIN AND CUTIN 507 Vegetable cellulose membranes were studied (Frey-Wyssling, 1 941) with the object of discovering whether in laminated systems the individual layers are of parallel texture, or whether it is a matter of interweaving. We have examples, such as the algae \^alonia (Van Iterson, 1933; Preston, Nicolai, Reed and Millard, 1948) and Chaetotnorpha (Nicolai and Frey-Wyssling, 1938), the laminated cell walls ot which can be split up into single lamellae of a few tenths of a // in thickness ; these lamellae are made up of strictly parallel fibrillae, which accounts for their striking cleavability parallel to the fibre direction. In consecutive lamellae the fibre directions cross at approxi- mately right angles (in 'Valonia at 78°); consequently the optical anisotropy of the individual lamellae is to a large extent mutually neutralized and, in transmitted light, the appearance is roughly that ofstatistically isotropic packets of layers. (Cf. Preston, 1947; Picken, Pryor and Swann, 1947). As opposed to these systems of membranes with uniform parallel texture of the individual lamellae, we have the fine-structure of the primary wall of cotton fibres. This thin membrane exhibits, according to Anderson and Kerr (1938), three different systems of striations, one of which runs perpendicular to the fibre axis, the two others falhng symmetrically at an angle of about 30° obliquely from the left and right. As the membrane cannot in this case be split up into three lamellae, presumably there are three different fibrillar directions in one and the same lamella. It may be supposed that submicroscopic fibriUae are interwoven in the three directions after the manner of a textile fabric. The observations made by Rosin (1946) on the tails of tadpoles would support the latter possibility. Judging by the arrangement of the pigment cells, which rest on a basal membrane of connective tissue, it would seem that the intrinsic texture of this membrane must consist of orthogonally trellised submicroscopic fibrillae of collagen. As it cannot be split up into two lamellae, the two systems of fibrillae apparently lie in the same plane. Rosin was able to show how the orthogonal fibrillar system grows by "afline" enlargement of the surface, the trellising of the two fibrillar systems always remaining rectangular (Fig. 150). Intussusception is responsible for surface enlargement, inasmuch as new submicroscopic fibrils are embedded in parallel. 30,8 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III In nature, therefore, there are very probably crossed micro- structural systems, the fibrillar structural elements of which interweave orthogonally or at other angles. The establishment of this fact may assist very materially in clarifying the submicroscopic texture of ex- ceedingly thin membranes. Fig. 150. Affine growth (from Rosin, 1946). a may grow larger progressively in pro- portion to h, or towards one side {c,d,e) ; the crossed system thereby remaining orthogonal. It is satisfactory that these speculations can be substantiated by electron microscopy. In primary cell walls (Frey-Wyssling, Muhle- THALER and Wyckoff, 1948) and in the Tunicata mantle (Frey- Wyssling and Frey, 195 i) interwoven crossed microfibrils are visible (Fig. 86d, p. 128). Therefore, in one and the same microscopic lamella, fibrillar elements may be laid down in different directions and, if they cross at 90°, their optical anisotropy is mutually cancelled when polarized light is transmitted perpendicularly. Rodlet and intrinsic birefringence of the chi tin frame. Chi tin was the first biological object in which the interaction of textural and intrinsic birefringence — discovered by Ambronn in artificial gels — could be demonstrated (Mohring, 1922). When decalcified specimens of chitin (lobster shell or lobster tendons) are immersed in solutions of potassium mercuric iodide of increasing refractive power, the bire- fringence decreases, falls to nil, changes its sign, reaches a minimum in the negative region, becomes nil a second time and then returns to positive (Fig. 151). The inference from this is that chitin is marked by a pronounced positive form anisotropy, i.e. rodlet birefringence, and a sHghtly negative intrinsic birefringence. In his imbibition experiments Castle (1936) finds reversal of the birefringence with mercuric iodide of potassium and iodobenzene in xylene, but not with other organic liquids (methylene iodide in CARBOHYDRATES, CHITIN AND CUTIN 309 0.012 0.010 0.008 0.006 0.004 0.002 0 -0.002 -0.004 \ \ \ \ •^ \ \ \ k \ \, V ^- \ - — ■ ^ 137 141 145 149 153 1.57 1.61 IdSn^ xylene, iodobenzene in alcohol). From this he concludes that the source of the negative birefringence is not natural chitin, but chitin chancred chemically by, say, potassium mercuric iodide. This con- clusion is, however, incorrect, for Diehl and Van Iterson (1955) found with mixtures of glycerol and quinoline, and Schmidt (1936) with a-monobromo-naphthalene (mixed with xylene) negative minima of the rodlet birefringence curve, even though these curves are not identical for various imbibition mixtures (Fig. 1 5 1). What was demonstrated in cellulose (Frey-Wyss- LiNG, 1936b) probably applies here, viz., that the difference in the adsorptive power of the micellar frame with respect to the components of the imbibition li- quid is responsible for the displacement of the curves. From the data now available, therefore, it may confidently be asserted that the submicroscopic chitin rodlets have a negative intrinsic birefringence. Cellulose likewise becomes optically negative by nitration and complete acetyl- ation (triacetyl cellulose), i.e., by the esterification of the polar OH- groups. It may therefore reasonably be presumed that it is the acetyl side chains of the chitin which cause the negative birefringence. It is nevertheless a curious fact that for chitin only one acetyl group per glucose residue is required for this, whereas three are necessary in cellulose; presumably, therefore, the amino group of the glucosamine also tends to produce negative birefringence. The negative intrinsic birefringence of chitin does not hamper the approach to the micellar texture of chitinous composite bodies by polarizing optics if the imbibition agents used are Hquids whose refractive index is below 1.48, i.e., water or glycerol. It should, however, be realized that the determination of the micellar orientation does not then take place on the basis of the positive intrinsic aniso- tropy of the submicroscopic frame of the membrane, as in cellulose, but rests on the positive rodlet birefringence of the chitin skeleton. Another interesting fact has been discovered based on the optical Fig. 15 1. Rodlet birefringence curve of chitin sinews (from Diehl and Van Iterson, 1935); A with quinoline-glycerol, B with mer- curic iodide of potassium. Ab- scissa: Refractive index n^ of the imbibition liquid. Ordinate: Bi- refringence n^-no. 3IO FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III properties of chitin. Young Lepidopferan scales yield a curve of rodlet birefringence with a minimum at ng = 1.57 instead of 1.61 (Fig. 151). From this behaviour Picken (1949) concluded that growing scales are not yet chitinous, and he proved that they consist at this stage of protein similar to muscle protein (n = 1.57) or keratin (n = 1.55)- No intrinsic double refraction is then visible ; it appears only in fully grown scales after the formation of chitin with its typical negative character. d. Starch Grains (Amylose and Amylopectin) Molecular structure of starch. The reserve carbohydrates sucrose, maltose and starch are a-glucosides (see page 60), in contradis- tinction to the skeletal carbohydrates cellulose, xylan, etc. which are jS-glucosidic. Compared to the straight cellulose chains, the glucosan chains with a- 1-4 bonds are rather kinked (Meyer and Mark, 1930). The result is that a spatial lattice of such chains must be less compact and, therefore, is more soluble, as indeed its physiological function as a reserve material requires it to be. Evidently the voids formed by this particular molecular configuration are partly filled with water molecules. Even the simplest a-glucoside, maltose, crystallizes with water of crystallization, and loosely bound water molecules also play an important part in the crystal lattice of starch. They do not, ad- mittedly, escape from the lattice as easily as from protein crystals, but when grains of starch are crushed, their lattice structure is likewise wrecked as the result of loss of water; they become amorphous, the birefringence and their X-ray diagram (Sponsler, 1922) vanishing. Hence additive water molecules apparently stabilize the lattice order of starch, as is the case in the reserve proteins. The chemistry of starch is complicated by the presence in the starch grains of two chemically distinct substances, viz., amylose and amylo- pectin. Amylose is soluble in hot water and is stained blue by iodine, whereas amylopectin swells in boiling water and gives a violet iodine coloration. Thus, when the starch grains 'form into a paste, amylose goes into solution, while the amylopectin becomes a swollen, in- soluble jelly. Neither component exhibits any reducing power upon Fehling's solution, which signifies that neither contains free aldehyde groups. K. H. Meyer (1940b) has discovered the difference in con- stitution between amyloses and amylopectin. He states that amyloses CARBOHYDRATES, CHITIN AND CUTIN 311 consist of unbranched chains, whereas amylopectin is made up of branched chains (Staudinger and Husemann, 1937; Staudinger, 1937 b) which together form a gel framework (Fig. 152); consequently amyloses are soluble, which amylopectin is not. Their other properties are given in Table XXVIII. In different starch samples the content of ^P?P^$:R>R?Ri3^PR? b) c) Fig. 152. Diagram of the molecular shapes of starch molecules. Glucose residues repre- sented as small rings: they are far more numerous (degree of polymerization) than shown here, a) Expanded amylose chain (cf. cellulose), b) Amylopectin (from K. H. Meyer, 1943): branched chain molecule. At x signs of the activity of the sugar-forming amylase: splitting off of the disaccharide maltose. In the absence of the dextrin-forming amylase, ■degradation ceases if maltose has split off from all the free terminals up to the branching place, c) Glycogen: highly branched starch molecule. TABLE XXVIII COMPARISON BETWEEN AMYLOSES AND AMYLOPECTIN Amyloses Amylopectin Molecular configuration Molecular weight (osmot.) j3-Amylase Pasting Films Tetramethyl glucose from maize starch Unbranched chain 1 000c - 1 00000 Complete hydrolysis Forms no paste Solid film 0.51% almost nil branches/molecule Branched molecule 5 0000 - 1 000000 Malto dextrin Forms paste Friable film 3-7% about 100 branches/molecule amylose varies from 34% to 0% (Table XXIX). The blue starch reaction with iodine is limited to amyloses with crystallized chains, i.e., unbranched chains orientated in parallel (Meyer and Bernfeld, 1941a), or to individual amylose chains wound up into a helix. 312 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III TABLE XXIX AMYLOSE CONTENT OF STARCH (from bates, FRENCH AND RUNDLE, 1 943) Starch % Amylose Ketan {Ory^a sativa f. glutimsa) . Waxy Corn {Zea mays f. saccharatd) Tapioca {Manihot utilissimd) . . Rice {Oryxa sativa) Banana {Musa sap/enfum) . . . . Corn (Zea mays) Potato {Solanum tuberosum) . . . Wheat {JTriticum aestivum) . . . Sago {Mefroxylon spec.) .... Lily bulb {JLilium spec.) .... o o 17 17 20.5 21 22 24 27 34 Freudenberg, Schaaf, Dumpert and Ploetz (1939) as also RuNDLE and Edwards (1943) argue that the chains of dissolved and precipitated amylose molecules are spirals, with six successive glucose rings to one revolution. Just as there are H-bonds between the neigh- bouring chain molecules of cellulose, so might there also be H-bonds between neighbouring turns of the same chain in the spiral model of the starch molecule. The six glucose rings per revolution can be compared with Schardinger's dextrins^, the molecules of which contain six to seven glucose residues (Hanes, 1957). Then, the inside of the hollow cylinders formed by the spiral chain provides the necessary space for the infiltration of iodine causing the blue starch reaction. Dextrins obtained as de- gradation products in starch hydrolysis give no iodine colour reaction when they contain only six or fewer glucose units. Dextrins con- taining eight to twelve glu- cose units produce red ra- ther than blue complexes. Only the longer amylose chains give the typical blue iodine colour. It is believed that the \ molecules are arranged along the centre of the amylose helix (Fig. 153). ^ Kratky and Schneidmesser (1938). Fig. 153. Model of iodine-filled amylose helix, (from RuNDLE, Foster and Baldwin, 1944). CARBOHYDRATES, CHITIN AND CUTIN 513 Molecules with branched chains produce red (amylopectin) or even brown colouring (glycogen) with iodine. The branching is due to glucosidic bonds from the aldehyde group of one amylose chain to another chain (Fig. 154). Such bifurcations CH.OH r -0, ' CHfiH non-reducing end groups branching member additional aldehyde branching end groups Fig. 154. End branching of amylopectin (from Frey-Wyssling, 1948 c). are frequently repeated and it can be shown that most of them corre- spond to 1-6 bonds (Myrback, 1938; Gibbons and Boissonnas, 1950). When branched chains are methylated and then hydrolyzed (Irvine, 1932), considerable quantities of dimethyl glucose (from branching junctions) and tetramethyl glucose (from the end members of the side chains) are formed in addition to trimethyl glucose. For amylopectin the amount of 2, 3 -dimethyl glucose formed is similar to that of 2,3,4, 6-tetramethyl glucose (5-5%). This means that the number of end members /// is about the same as the number of bifur- cations b. A dichotomous branching would satisfy this relation, because it yields h ^ m — i (Fig. 155). aldehydic end member 1 bifurcation 2 bifurcations 16 end members 64 end members (b) Fig. 155. Dichotomous branching of amylopectin, a) in a plane, b) in space (from Frey- Wyssling, 1948 c). 314 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III For amylopectin of Ys million molecular weight (Table XXVIII, p. 311), which corresponds to a degree of polymerization of 2000, 64 bifurcations and as many end members would be found. The average length of the branches with 1 5 glucose units would measure about 50 A, so that the brush-like molecule of Fig. 155b would cover 350 A in an expanded state. The amylopectin content varies in the different kinds of starch, which accounts for the familiar specific differences between them. Potato starch has a higher amylopectin content than wheat starch (Meyer and Bernfeld, 1941b). Ketan, the starch grains of which are dyed red bv iodine, contains only amylopectin of high molecular weight (Meyer and Heinrich, 1942). Amylopectin possesses weakly acid properties and can therefore be separated by electrophoresis from amyloses, which are absolutely neutral (Lamm, 1937). Presumably the acid groups in amylopectin are responsible for the fact that only basic dyes can stain starch grains with a colour which is fast to washing, Samec (1927) says they consist of phosphoric acid. Meyer and Mark questioned in 1950 the existence of phosphoric ester bridges between the glucose chains, and nowadays amylopectin is regarded as free from phosphoric acid (Meyer and Brentano, 1936; Samec, 1942), The discovery by Hanes (1940) that the enzymatic degradation of starch is a phosphorolysis, and not hydrolysis, invests the phosphorus content of starch grains with a particular significance. This knowledge led to the synthesis of starch in vitro. Starting from phosphorylized glucose, Hanes united it with the enzyme phosphorylase ; when equi- librium sets in between glucose- 1 -phosphate and starch, this com- pound, owing to its insolubility, is synthesized. Hanes' synthesis of starch is the first instance of an artificial manufacture of a high polymeric natural product. The decomposition of starch is a highly complicated process of fermentation (Myrback, 1938; Myrback and co-workers, 1942). Amylase, the enzyme which decomposes starch, consists of two different constituents, viz., the dextrinogenous a-amylase and the saccharogenous /9-amylase, both of which have been isolated and crystallized (Meyer, Fischer and Bernfeld, i 947 ; Meyer, Fischer and PiGUET, 195 1 ; Meyer, K, H,, 195 i). The latter splits off maltose (twin groups of glucose) from the aldehyde end of the starch chains (Fig. 152b, p. 311), but is unable to break up the branch junctions CARBOHYDRATES, CHITIN AND CUTIN 315 of the amylopectin. a-amylase is able to break down the amylopectin into soluble fragments (dextrins) without at first generating maltose; subsequently saccharification sets in by degrees (Meyer and Bern- FELD, 1941c). The branching junctions of the amylopectin are in- accessible to the /5-amylase, for, besides the usual 1-4 bond between the glucose residues, there is an additional 6-1 bond passing into the side chain, the splitting of which needs yet another enzyme. The microscopic structure of starch grains. The microscopic structure of starch grains has been dealt with so often and so exhaustively (see Badenhuizen's comprehensive review 1937) that, to avoid repetition, I shall here touch only on a few points which appear to me of par- ticular importance (Samec, 1942/43). The familiar arrangement in layers of starch grains is brought about, in the unanimous opinion of the majority of investigators from Nageli (1858) to our contemporaries, by alternate layers of stronger and weaker refractive power, or containing a smaller or larger percentage of water. Now if a weakly refractive, narrow layer were bordered both inside and outside by neighbours of higher refractivity, it would shine brightly in the microscope at low adjustment and, when the tube is raised, the bright Becke lines on both sides should pass over into the optically denser layers. This, however, does not take place (Frey- Wyssling, 1936a, page 287). With pronounced stratification, especially of eccentric starch grains immersed in water (potato, Pellionia, etc.), it may be seen distinctly how the Becke line at the edges of all layers moves outivards only when the tube is raised. True, there are cases when a pale lustre can be seen to shift inwards, but on the outside it is always incomparably stronger. From this we may confidently conclude that every layer is more refractive on the inside, the refractive power outwards as a rule diminishing quite gradually, and then suddenly coming up against a layer of higher refractive index. Thus, in a section through the grain the refractive power in the various layers is not equally high or low, but there is a continuous decrease towards the edges and discontinuous increase at the outer edges of the layers, as represented in diagram bv the 0 12 3 4 5 Fig. 156. Microscopic lamination of starch grains. Diagram of the refraction. Ab- scissa: 1-5 layer edges. Ordinate: Refractive index «. 3l6 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III serrated line in Fig. 1 5 6. The arrangement, therefore, of the layers is not dense/light/dense, but rather a gradual reduction in density in each layer up to the edge of the next layer, at which point the density suddenly rises again to its initial level. The fact noted by Young (1938) that the layers of the starch grains of Canna are more easily stained outside than inside is probably due to the looser structure of the outer portions of the layers. The inference from the foregoing as to the apposition growth of starch grains is that, in the formation of a new layer, the deposition is at first dense, becoming looser little by little until, at a given degree of impoverishment, growth ceases altogether. The sugars consumed have then probably to be made good before the process can start afresh. As Van de Sande Bakhuizen (1925) showed, if external con- ditions are constant, lamination does not occur, because nutritive material is then always available in the same concentration and, therefore, there is no impoverishment during growth. The same applies to the lamination of cotton fibres, which likewise depends largely upon external conditions of growth, viz., temperature (Kerr, 1937), or can, indeed, be prevented altogether by constant exposure to light and the exclusion of fluctuations in temperature (Anderson and Moore, 1937). During their entire growth, the starch grains are enclosed within the amyloplast, which produces them; this stretches very considerably in the process and finally becomes an exceedingly thin, scarcely perceptible pellicle enveloping the grain. Radial structures have for long been observed in addition to lami- nation ; they take the form of corroded patterns during the mobili- zation of the starch in the germinating seeds, or of thin radial cracks. The starch grains have therefore been thought to be of spherite tex- ture. This would seem to receive support from the optical fact that a positive spherite cross always occurs (see Fig. 66, p. 96), because the starch chains, like cellulose chains, are optically positive as referred to their long axis. The starch grains can be split up tangentially and radially by chemi- cal means into minute blocks of i ^ edge length (Hanson and Katz, 1934; Badenhuizen, 1937) and these particles have been said to be pre-formed elementary units of the starch grain (Fig. 157). Structures of the kind are obtained if starch granules are treated for days with 7^4% hydrochloric acid and are then swollen in 2 molar Ca(N03)2 I CARBOHYDRATES, CHITIN AND CUTIN 517 (known as "Lintnerization"). Hanson and Katz suppose that the blocks consist of packets of amyloses and that the swollen inter- mediate substance is amylopectin. This view is not borne out by the staining properties of the substances, since the basic dyestufF, fuchsine, stains the blocks a deep red, whereas rvOOCZl^^ the supposed amylopectin intermed- r^P*^ nn^'^'^Q^ iate substance remains colourless. r^rr€^^^:^rf4=^lf]^^^ Seeing that this block structure is c96^<^K^^°^^^m> ^S\ formed only after the application of 8§Hmo^^o\P§^ strong hydrolysis with hydrochloric RR QQfx'^^ °n! )o^ '^'^ acid, the view of pre-formation is hardly tenable; it is more likely to be a case of hydrolysis patterns (Frey- Wyssling, 1936a, p. 290). This kind of partitioning of objects made up of high-polymeric chain molecules has Fig. 157. Microscopic block structure likewise been observed in theproduc- "^ "linjnerized" wheat starch (from ^ Hanson and Katz, 1934). tion of chemical cross-sections through cellulose fibres with sulphuric acid (Kelaney and Searle, 1930), the decomposition of cotton into "dermatosomes" by hydrochloric acid (Farr and Eckerson, 1934) and m the decomposition of muscle fibres in acid alcohol (Schmidt, 1937a, p. 180). Thus in all these cases hydrolytic agents are necessary to produce the reported dissociations. Considering how sensitive high-polymeric main valence chains with glucoside or peptide bonds are to hydrolysis, it is out of the question that the reagents used would merely have a dissolving effect; thev surely cannot fail to induce break-down and decomposition. Cellulose chains, for instance, are broken down by i n HCl (= 3.6%) at 53° C. in six hours from 1660 to 445 degree of polymerization (Staudinger and Sorkin, 1937a) and amylose chains are shortened in only 3^/, minutes by 2 n HCl from 940 degree of polymerization to one-fifth their length (Staudinger and Husemann, 1937). Nor is it surprising that this hydrolytic degradation should take place mainly across the particular texture (fibrous or spherite), since the hydrolysis occurs perpendicularly to the alignment of the thread molecules. The par- titioning parallel to the axis of orientation need not necessarily be of a hydrolytic nature; it is as likely to take place in a less drastic, physical way (radial cleavage, cracks due to drying, fibrillation through 3l8 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III swelling), for in this direction there are chiefly secondary and not primary valencies that have to be overcome. A question which it is difficult to answer is why the hydrolysis should occur with such characteristic rhythm in the case of these microscopically homogeneous structures. If the microscopic segments which are formed correspond to the chain lengths of the macro- molecules, as Staudinger, Staudinger, and Sauter (1937) assume that they do in the laminate break-down of synthetically produced polyoxymethylene crystals, a mechanical cleavage perpendicularly to the crystal axis should occur; but starch molecules are not of microscopic length. Therefore, any such interpretation would not apply to starch grains. Another possible explanation is that maybe the submicroscopic capillary system of the object in question, corre- sponding to the hydrolysis pattern, is periodically fine and coarse. Without any such auxiHary hypotheses, however, it is possible to suppose that in the hydrolytic break-down of fibrous or spherite structures, fragments of uniform size are produced, just as, in the mechanical pulverization of crystals or glass, only segments or splinters of approximately the same size split off, this size having nothing to do with the structural elementary units, but depending solely upon the method of comminution applied. Macroscopically as well, objects of entirely uniform structure can be split into pieces of similar size which have not been pre-formed; thus, when ice is broken up, a perfectly homogeneous slab of ice may split up into floes of equal proportions, the size of which is by no means predetermined. Under certain circumstances and, of course, to an enormously en- larged scale, the pattern of the floes may be strikingly reminiscent of the block structure represented in Fig. 157. In the opinion of Baden- HUiZEN (1938) the "blocks" certainly are not pre-formed in the structure of the starch grains. The submicroscopic structure of starch grains. Katz and Derksen (1933) have established that different kinds of starch do not produce the same X-ray spectrum. For example, the gramineous starch of wheat, rice, corn and oats produces what is known as an A spectrum, whereas potato starch has a B spectrum, and both, when formed into a paste, produce a third, called the V spectrum. Starches with a B spectrum have been converted at higher temperatures to the A kind (Katz and Derksen, 1933); it has also been shown that the V spec- I CARBOHYDRATES, CHITIN AND CUTIN 519 trum reverts to a B spectrum in the so-called retrogradation of paste, in which process the quantity of bound water plays a certain part. Thus the following conversions may be observed in wheat starch which is pasted up and then retrogresses: A ^ V ^ B. Several investigators (Sponsler, 1923; v. Naray-Szabo, 1928) have attempted to deduce the size of the elementary cell of crystallized starch. Bear and French (1941) find for B starch an orthorhombic cell with a volume of 930 A^ and for A starch a triclinic cell with 843 A^ volume. This is much more than the cellulose cell which occupies only 670 A^. This proves that, besides glucose residues, water molecules are enclosed in the cell. But these results are doubtful, as starch produces only powder diagrams, i.e., Debye-Scherrer rings. Recently Kreger (1946, 195 1) has succeeded in irradiating only part of the large starch grains oi Phajus grandifolius by a special micro- method. In this way he gets a fibre pattern, which enables him to calculate the cell of B starch more exactly. Rundle, Daasch and French (1944) were able to prepare artificial amylose threads, which yielded a fibre period of 10.6 A, whereas that of cellulose is only 10.3 A. They think that the two glucose residues of the glucosan chain is somewhat stretched in crystallized B-starch, whereas Kreger (195 1) places three helically arranged a-glucose rings into the distance of 10.6 A. When the results of the investigators mentioned are combined, the following orthorhombic unit cell is found for crystal- lized starch (B-diagram) : a : b : c = 9.0 : 10.6 : 1 5.6 A. Of these periods a : c show the ratio i : -\/5, indicating a hexagonal symmetry. This is in accordance with a threefold screw axis along the chains suggested above. The hexagonal unit cell has the periods a : b = 18 : 10.6 A and contains 18 chains, i.e. 54 glucose residues and 54 water molecules. The density of starch under water of 1.60-1,63 is in agreement with this unit cell which is illustrated by Fig. 158 showing two possible arrangements of the starch chains (Kreger, 1951). Senti and Witnauer (1946) have shown that in the A spectrum of starch there is also a fibre period of 10.6 A. From this it follows that in the starch grain of either A or B type the amylose chains are expanding, forming a spiral pitch of 10.6 A with three glucose 320 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III residues. Consequently it is only in solution that they contract to a pitch of 7.8 A (RuNDLE and Edwards, 1943) formed by six glucose residues. It is obviously misleading when, in analogy to the proteins, the extended amylose chain is called "denatured" amylose, because its natural state in the starch grain seems to be the expanded modi- fication. «> / ._; \ 1 / 1 \ / • ' • \!/ • ' • \ Fig. 158. Cross-sections of the two possible unit cells of crystallized starch (from Kreger, 1951). The optical behaviour of the starch grains rules out contracted chams in their structure (Frey-Wyssling, 1940 c). Since the con- tracted amylose chains show their highest polarizability (corre- sponding to n^) perpendicular to the helical axis which, according to X-ray evidence, runs radially, the starch grains ought to be optically negative. But, as mentioned above, they represent optically positive spherites. Therefore, they must contain expanded chains which have their highest polarizability (n^,) parallel to the fibre axis. The iodine dichroism points in the same direction. Contracted amylose chains (Fig. 1 5 3, p. 312) have their highest absorption coefficient (k^) parallel to the fibre axis, which consequently runs perpendicular to n^. In expanded chains, however, k^ and n,, coincide. Since this coincidence is characteristic for starch grains, it must be taken for granted that they consist of expanded chains. Further proof of radially orientated elements is provided by the existence of rodlet birefringence in starch grains (Speich, 1941). This raises the question as to what type of submicroscopic spaces permits the penetration of imbibition liquids. They cannot have the same character as in cellulose, because hitherto no submicroscopic struc- CARBOHYDRATES, CHITIN AND CUTIN 321 0.074 0.013 tural elements of starch have been found with the electron microscope (unpublished data) so that its texture must be amicroscopic. On the other hand, those spaces must be fairly wide, since they are accessible to the big molecule I.^ whose diameter measures 2.7 and 5.3 A, and to organic dyestuffs. It is doubtful whether these molecules can be inserted into the intermolecular spaces of the crystal lattice shown in Fig. 158, where the cross-section of the starch chains is not drawn in its actual dimensions, but is merely symbolized by a black dot. However, there is the possibility of gaps in the chain lattice, or of a widening of the crystal lattice by water, in the same way as is known to occur in zeolites. The analysis of the rodlet birefringence supports this view\ Series of aldehydes, monovalent alcohols and polyvalent alcohols (glycol, glycerol) including wa- ter give three different curves of rodlet birefringence (Fig. 159), disclosing different intrinsic double refractions in these three groups of liquids. This is due to the different interaction of these compounds with the starch chains. Lipophilic liquids (amyl- bromide, xylene, toluene, ben- zene, chlorobenzene, bromoben- zene, and a-bromonaphthalene) do not penetrate into the starch grain, because they have no affin- itv for starch ; the double refrac- tion therefore does not change in a series of lipophilic liquids with increasing refractive power. Since crystallized starch attracts water molecules which penetrate between the molecular chains, the starch grains ought to dissolve in water. Such a dissolution is possible with another reserve carbohydrate, viz., mannan, which is obtained from the tuber of Amorphophallus konjak and is 0012 OOII 1.60 ni" Fig. 159. Rodlet birefringence curves of potato starch (Speich, 1941). A in lipophilic liquids; B in aldehydes; C in monovalent alcohols (except ethanol); D in water, ethanol, glycol, glycerol and their mixtures. 322 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III marketed in Japan under the name of "Konyaku". The technical com- mercial product consists of irregularly bordered granules which light up in a quite irregular manner between crossed nicols and reveal no ordered structure (Fig. i6oa). But when these granules of konyak are observed in water, they are seen to undergo a remarkable change. Under the very eyes of the observer, they swell and assume a spherotexture, exhibiting a positive a) b) Fig. 1 60. Alannan granules of konyaku {Amorphophallns konjak) in the polarizing micro- scope. Embedding medium a) xylene; h) water; there is the transient appearance of a spherite cross. spherite cross (Fig. i6ob). After a time the appearance becomes fainter and eventually vanishes altogether, because the swelling is not limited, but continues until solution takes place. This phenomenon may be interpreted as follows: The mannan chains, which are comparable to the amylose chains, in the dry konyak granule (which it is best to examine in a hydrophobic embedding medium, such ac xylene or Canada balsam) are arranged in bundles that accumulate in the granule without any defined orientation. This explains why the polar- izing picture is irregular. As soon as water is added to these chains, however, they are hydrated, become mutually mobile and align themselves radially and this results in a spherotexture. The conditions are much the same as those in myelin figures (see Fig. 47, p. 56), except that hydration is not limited to one hydrophilic pole of the molecule, but encompasses the entire mannan chain and ultimately spreads to such an extent that the individvial thread molecules become independent of each other and go into solution. CARBOHYDRATES, CHITIN AND CUTIN 3^5 Fig. i6i. Diagram of the submicro- scopic structure of a layer of a starch grain, a Outer portion looser, less refractive, with little interlinking; / inner portion denser, more refractive, more closely interlinked. The stage of voluntary spherite formation is comparable to the structure of starch grains. The starch molecules are obviously also fixed in a radial direction by water of hydration. In this case, however, the water does not function as a solvent, but participates, as a loosely bound constituent, in the build-up of the spatial lattice. The starch chains are far more highly polymeric than the molecules of konyak mannans . Notwithstanding this, the amy loses are soluble and, if the starch grain seems nevertheless to swell only to a limited degree, there must be some particular hindrance to solution. This is probably to be found in the amylo- pectin, the glucosan chains of which are interlinked. There is good reason to believe that these amylopectinous linkages occur in each individual layer of the starch grains in the inner, denser and more refractive portions and that independent, amylose chains are accumulated in the outer, looser portions of the layers. Jaloveczky (1942) states that the lamellae containing amvlopectin are isotropic, whereas those containing amylose are anisotropic and can be stained. It has been suggested that all the amylopectin is localized in the outermost marginal layer of starch grains, which is resistant when they are made into paste. It would seem more probable, however, that the starch is liable to every conceivable transition from the easily soluble amyloses to the virtually insoluble constituent of the amylopectin, which resists even enzymatic degradation. Thus it might be supposed that amylopectins occur in the denser portions of all the layers, though not in the same degree as in the insoluble outside layer which resists when starch is made into paste. On this assumption the submicro- scopic structure of a starch grain layer was represented in 1938 as in Fig. 161. This diagram takes into account the following observed facts : The density and refractive index at the core of a layer diminish gradually towards the outer regions and then increase suddenly at the boundary of the layer. The solubilitv is not equal everywhere within the layer. The water of constitution between the chains is partially bound as 324 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III hydration water of the lattice and is partially mobile as swelling water. The swelling maximum of the grains is governed by the linkage of the chains. Adjacent layers have coalesced. The structure is wide- meshed and porous, causing colourability and rodlet birefringence (Speich, 1 941). In the process of pasting, the loosely linked, or Fig. 162. Possibilities of fine-structure in a layer of a starch grain (n = degree of poly- meri2ation). a) Fine-structure of amy lose (n'-^2 5o); />) fine-structure of amylopectin (n'-^8ooo); c) mixture of amylose {n^^z'^o) and amylopectin( n -^ 2,000,000); d) amylopectin with inward pointing aldehyde group ; e) amylopectin with outward pointing aldehyde group; /) amylopectin molecules with opposite orientation (from Frey- Wyssling, 1948 c). unlinked glucosan chains go into solution as amylose, whereas the strongly linked amylopectin chains agglutinate throughout the paste. Minor specific or individual variations in linkage may be responsible for the peculiar resistance of different kinds of starch, or of different grains within the same kind of starch. For instance, there are grains of potato starch, the peripheral layer of which is so resistant to ■enzymes that some of them may pass unaffected through the intestines (Weichsel, 1936). To-day a more detailed discussion of the fine-structure in the starch grain is possible, because the chemical constitution of amyloses and amylopectins has been cleared up since 1938. Fig. 162 shows some possible arrangements of these molecules with different degrees of polymerization n in a layer i ^ thick of a starch grain (Frey-Wyssling, 1948c). The simplest case is represented by a), where only amylose molecules with n--^ 250 (0.088 ix length) are drawn. However, since the amyloses form only a minor portion of the starch grains (Table CARBOHYDRATES, CHITIN AND CUTIN 325 XXIX, p, 5 1 2), the arrangement of the amylopectins is more important, b) and c) show such molecules of n '-^ 8000 and n -^ 2,000,000; it is remarkable that a dichotomous amylopectin molecule of n ~' 2,000,000 should have only the same length {ca. 0.09 ;*) as an amylose molecule of n/--' 250. The many end members of the amylopectin molecule are Fig. 163. Interpenetration of amylopectin chains of opposite orientation (from Frey- Wyssling, 1948 c). Arrows indicate the non-reducing end of the chains. Dotted areas = crystalline regions (cf. Fig. iGzf). not aldehyde in character; only the glucose residue at the starting point of the bifurcated high polymer has an open aldehyde group. Therefore, it is likely that such a molecule grows by adding new glucose molecules with the active aldehyde group to the brush end. This is the reason why in Fig. 162 the molecules have been oriented in such a way that their growth direction coincides with that of the apposition growth of the starch grain. Since the amylopectin is attacked by the /5-amylase from the non-aldehydic end, this would explain why the enzymatic dissolution of the starch grain often starts at the outer boundary of its layers. However, such an arrangement would cause a higher density of the layer in its outer portion as seen 326 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III from d), and this is in contradiction to the optical result which proves the outer portion of the layers to be less dense than their inner portion (Fig. 1 5 6, p. 515). Therefore, an arrangement as indicated in e) would better correspond to the optical behaviour of the grains. But then the inner portion of the layers ought to be attacked first by ^-amylase. This contradiction and the fact that no chemical polarity of the layers has ever been observed, make a compromise probable as shown in f). If the amylopectin molecules grow in both directions, the layer will be chemically uniform. Further, branches running in opposite directions may crystallize with each other (Fig. 163). Since in the crystal lattice of cellulose the glucosan chains run also in opposite directions, such a structure for starch is quite probable. The diagram of Fig. 163 would allow of a mixed crystallization of amylose with amylopectin and it shows how gaps may arise in the crystal lattice of starch. Since the X-ray diagram is that of a fibre texture, the two directions of the bifurcating chains cannot be crossed as in Fig. 163, but must run almost parallel. Of all the theories so far developed for the structure of starch grains, that propounded by A. Meyer (1895) comes nearest to the views set forth here. Instead of his dendritic branching, however, we assume that there is all-round interlinking, and that the dimensions of the structure are reduced by some orders of magnitude to the molecular. § 2. Proteins a. Reserve Protein There is a fundamental difference between reserve proteins and fibrous proteins. First and foremost, the reserve proteins are soluble in water, dilute salt solutions, acids and alkalies, whereas the distin- guishing feature of the frame substances is their pronounced in- solubility. Reserve proteins frequently tend to crystallize if the solvent is withdrawn in the proper way, as, for instance, by natural means in the formation of aleurone granules owing to the drying up of vacuoles in vegetable storage tissues. Polyhedral, crystallized corpuscles are then formed, different, however, from real crystals in that they are liable to swell and to take up stain. Nageli (1862) therefore called them crystalloids. Notwithstanding the fact that the term "crystalloid" was later applied by Graham in quite another, and etymologically 2 PROTEINS 327 incorrect, sense to real solutions of substances of low molecular weight, Nageli's original definition was retained by botanical cvtologists, for to this very day the enclosures of the aleurone granules in the seeds of Ricinns (Pfeffer, 1872), Momordica (Zimmermann, 1922), Telfairia (Leuthold, 1935), etc. are called crystalloids. The crystal lattice of globular proteins are often cubic or hexagonal ; witness the occurrence of cubic or rhombohedral crystal shapes in the crystallized reserve proteins of vegetable seeds. The globular ele- mentary units of the molecular lattice (see p. 26) are so big as to produce a large spaced lattice (Fig. 90a, p. 136), into the meshes of which swelling agents and dyestuffs can penetrate. The swelling of the rhombohedral protein crystalloids is anisotropic, being, as Nageli (1862) had already discovered, different parallel to the crystal axis from what it is perpendicular to it. Up to 1939 only seven of all the many crystallizing globular proteins had been examined by X-ray crystallography, these being pepsin, insulin, excelsin, lactoglobulin, haemoglobulin, chymotrypsin and tobacco seed globulin (Crowfoot, 1939, 1941). For, in spite of repeated attempts, it was long before any success crowned the efforts to obtain X-ray photographs of mono- crystals of the crystalloids. Thus, for example, the crystalloids of the seed globulin excelsin of the spruce have threefold symmetry, and those of pepsin hexagonal, but this fact was in no way revealed by the X-ray photograph of a single crystal. On the contrary, until a short while ago all monocrystal photographs of globular proteins, and particularly in the case of the well "crystallized" pepsin (Astbury and LoMAx, 1934), only produced Debye-Scherrer rings with lattice spacings of 4.6 and 11.5 A, which unexpectedly proved to have the backbone thickness and the side chain spacing of polypeptide chains. In view of the large molecular weight of the crystallized proteins, it was anticipated that, instead of such spacings, there would be very large periods which would produce interference dots quite near the centre of the photograph. Although some such large lattice spacings had been found in insulin (Clark and Corrigan, 1932) and in pepsin (Fankuchen, 1934), Bernal and Crowfoot (1934) were the first to be entirely successful in obtaining monocrystal X-ray diagrams. The secret of their success lay in the fact that they irradiated the pepsin crystalloids (hexagonal bipyramids 2 mm in height) in their mother liquor. In this way they discovered a wide-meshed crystal lattice, the 3z8 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES nr elementary regions of which harbour globular macromolecules of about 40,000 molecular weight, a figure that tallies with the values found in the ultracentrifuge. Fig. 164 shows the result of such an investigation of crystallized insuUn by Crowfoot (1938, 1941). It is characterized by contour lines Fig. 164. Patterson-Fourier diagram of crystallized insulin (from Crowfoot, 1938). in the unit cell of the lattice which are derived from intensity measure- ments of the X-ray diffraction pattern. The resulting so-called Patterson-Fourier diagram shows the trigonal symmetry of the crystal lattice in a most instructive way. The moment the crystalloids are removed from the mother liquor^ however, and are exposed to the air, they denature and produce only powder diagrams. Although they retain their crystallographic shape outwardly, apparently the internal regular crystal lattice order can only exist for just so long as the solvent is distributed between the macromolecules . It would seem that there is some relationship between globular reserve and fibrillar frame proteins, notwithstanding the great differences between them in point of solubility and the morphology of the molecular elementary units, for Astbury, Dickinson, and Bailey (1935) succeeded in producing filaments and films from the seed globulin edestin and from egg albumin which, when elongated,, exhibit the ^-keratin type of fibre diagram. Astbury therefore assumes the presence of folded polypeptide chains in the crystalloids of the reserve proteins, as represented in Fig. 165. In this way certain self- contained isodiametric areas might be imagined, corresponding ap- 2 PROTEINS 3^9 proximately to the globular molecules of the reserve proteins, but only capable of existence in equilibrium with molecules of the solvent. Where there is denaturation, these loosely-knit complexes would dis- sociate and long chains would begin to form across the intervening spaces. This would explain why denatured reserve proteins become NH. ,C0 — -NH-CHR-COy^ iNH-CHR-CO-NH CO-NH-CHR-CO ^NH-CHR-CO- — CO-CHR-NH' *C0-CHR-NH-C0 NH-Cn-CHR-NH ,CO-CHR-NH— —NH-CHR-CO^ ,NH-CHR-CO-NH CO-NH-CHR-CO ^NH-CHR-CO— — NH-CHR-NH' ^CO-CHR-NH-CO NH-CO-CHR-NH. CO-CHR-NH— CO *NH Fig. 165. Molecular structure of a protein crystalloid. The arrows mark the bonds which, in "degeneration" to a fibre protein, are resolved to form bridges over the intermediate spaces (which contain solvent) to the neighbouring molecules, by which means straight chains come into existence (from Mark and Philipp, 1937). less digestible, since in this process the polypeptide compounds pass from a loosened soluble form to the insoluble chain lattice form ,of the frame-protein type (see Fig. 90, p. 136). Miss Wrinch (1937) suggests that ring formation of polypeptide chains may be responsible for the globular shape of reserve protein molecules. According to her "cyclol theory", the chains would form hexagons by ring folding and forming a bridge bond at the open position between the NH and CO groups. If, by tTiis scheme, six amino acids are assigned to a cyclol six-ring, the result is three regular hexagons arranged trigyrically around a central hexagon. This ternate arrangement falls into line with the trigonal or hexagonal crystal system of the crystallized reserve proteins. For each bridge formed, an alcoholic C(OH) group comes into existence (see page 158), all the hydroxyls of which lie on the same side of the ring system; this will therefore have a hydrophilic and a hydrophobic side and there will thus be a tendency towards double layer formation. On this view, the protein crystalloids are to be conceived as packets of double layers of this kind, the hydrophilic planes being responsible for the ability of the crystal lattice to swell. In recent years, it has become doubtful whether in globular proteins the peptide linkages characteristic of fibrillar proteins are already preformed (Jordan, 1947; Scheibe, 1948). Because of their 330 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III pronounced dipole character, there is a strong mutual attraction between amino acids. In +H3N.CH3.C02~ this leads to the formation of molecular layers when glycine crystallizes. For the other amino acids there is a steric hindrance due to the side chains R, which prevent the formation of closelv packed layers. Those amino acids therefore -T - -I X'-- Fig. 1 66 Fig. 167 I Fig. 166. Aggregation of amino acid dipoles by three causes threefold symmetry. Peptide bonds are not yet realized (from Scheibe, 1948). Fig. 167. Plate of 24 amino acid groups of three. Groups I (double rings) lie in a somewhat higher plane than groups II (simple rings) ; opposite borders of the hexagon differ from each other, causing polarity in the direction of the arrows. X' — Z cross-section (from Scheibe, 1948). associate in threes with the side chains R, pointing in three different directions in the plane in which -NH3+ and -COO" lie (Fig. 166). These groups of three attract each other, forming hexagonal rings. In such a ring three groups lie in a somewhat higher plane and three in a lower one (Fig. 167). The hexagonal rings represent a molecular layer with trigonal symmetry. These layers can be superimposed, yielding a hexagonal crystal lattice. The peculiarity of such a crystal is that it consists of amino acids which still retain their individuality and are not tied together to form polypeptide chains. It represents a "protein" without peptide bonds. Denaturation would then imply the formation of peptide bonds between adjacent amino acids. Arguments in favour of such a view PROTEINS 351 are these: There are globular proteins, such as haemoglobin, which are not attacked by polypeptidase until they are denatured (Hauro- wiTZ, 1949), but since many proteolytic enzymes work under con- ditions which cause denaturation (e.g. pepsin at p^ i), this fact is often obscured. Further, it seems that the three amino acids of the pre- formed groups in the molecular layer (Fig. 166) form tripeptides when denatured (tripeptols of Jordan, 1947). On the other hand, there is no indication of how such an arrange- ment leads to molecules with a definite weight. When 4 layers as seen in Fig. 167 with 24 x 3 amino acids are superposed, a molecule of 288 amino acids is obtained, which would correspond, for instance, to insulin. It is not clear, however, why piles of only four layers exist, and why aggregations of such fourfold layers by 2, 4 etc. occur according to the Svedberg series. It seems likely, therefore, that the binding forces inside the molecules are stronger than those which cause the aggregation of globular protein molecules to multiples and crystal lattices (Fig. 84, p. 126). b. Si/k (Silk Fibroin) Microscopic and suhmicroscopic structure. A cross-section of the cocoon thread of the silk-moth {Bombyx mori) reveals two halves in mirror symmetry, which owe their existence to the paired silk-glands. These ■produce two discrete fibroin threads which are covered with a layer Fig. 168. Fine-structure of silk, a) Microscopic cross-section through the cocoon filament (after Ohara, 1933a); b) suhmicroscopic structure of the fibroin thread, i Skin and 2 cortex (fibroid texture with tangential scattering) of the sericin layer. 3 Skin, 4 cortex (fibroid texture with radial scattering) and 5 central zone (fibrous texture) of the fibroin filament. of sericin (Fig. 168). The regular structure as seen in Fig. 168 a is apparently disturbed where the threads cross in the cocoon, which would go to show that the thread is still plastic when it leaves the 332 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III silk-gland. A finer structure is revealed both by the sericin layer and by the fibroin threads (Ohara, 1933 a). On the outside is a very weakly birefringent, almost amorphous membrane of sericin, under which comes a strongly birefringent layer of a fibroid texture. The sericin layer is separated from the fibroin threads by an isotropic lamella. It is here that the sericin becomes detached from the fibroin when the silk is degummed in a dilute soap solution. Two degummed silk threads are then formed from every cocoon filament or raw silk thread. The difference between raw silk and ordinary degummed silk is therefore that the former is still surrounded by the sericin cortex, though admittedly this often suffers considerable mechanical damage. The fibroin filaments, which are now to form our main topic, have, according to Ohara, three zones which are optically distinguishable, i.e., a central zone of fibrous texture, a cortical layer around this of fibroid texture and, finally, at the outside a skin layer. This is only slightly anisotropic, yet its texture is apparently slightly fibroid. It is interesting to note that here the prevailing direction of orientation — i.e., deviation of the optically positive submicroscopic fibroin rodlets — is not tangential, as in the sericin layer or in cellulose fibres, but radial (see Fig. i68b). It seems that in the process of degumming, the character of the scattering in the coating layer changes from radial to tangential, for, after the hot water treatment, the large axis of the index ellipse lies tangentially. The scattering of the fibroid texture of the cortical layer is likewise radial. Thus in a cross-section through the cocoon threads, sericin and fibroin are easily distinguishable by their different optical behaviour in a polarizing microscope in which a selenite test plate has been inserted, in that the sericin wrapping produces a negative, and the fibroin cortex a positive spherite cross (cf. Fig. 66a, b, p. 96). In the cross-section the central zone appears to be isotropic. The zoning of the fibroin filament is brought out clearly by dichroic gold and silver staining. In conformity with its fibrous texture, the central zone exhibits pronounced dichroism; in the cortical zone, on the other hand, with its far inferior orientation, the coloration is not dichroic, and in the coating layer there is none at all. According to Ohara (1933 a), this is how the fibroin filament laminates as a result of coagulation: The coating layer is the first to coagulate on leaving the silk-gland, before there is any opportunity for an ordered sub- 2 PROTEINS 333 microscopic structure to be formed. A little later, the cortical layer, the jfibroin thread molecules of which are already to some extent orientated, coagulates. The fibroin mass in the central zone remains plastic for a longer period and the chain molecules of the silk fibroin are all able to orientate with parallel axes before they combine to form •fs!e*-.«=5S«.-., ^^"^ ik 7> - -c? r v< a) b) Fig. 169. Fine-structure of silk (from Ohara, 1933b). a) Beading with Ca(N03)2 solution; b) fibrillar formation with hypobromite. a micellar frame. As rayon filaments often display a similar structure (Preston, 1933), Ohara's hypothesis is attractive, but it should be pointed out that, unlike natural silk, the cortical part of viscose is submicroscopically better orientated than the central part of the filament ("skin effect" according to Preston, 1933). In rayon, the stretching process brings about an orientation of the peripheral region, whereas the thread molecules of the still uncongealed mass in the centre of the filament are not effectively held by the orientating forces, owing to their mobility. Hence the optical conditions prevailing in natural silk which conflict with this interpretation must be explained in some other way. Since, like vegetable bast fibres, the silk fibroin filaments possess a central portion of a fibrous texture and a skin with a pronounced deviation of the microfibrils from the direction of the fibre axis, their swelling and hydrolysis phenomena are similar to the cellulose walls ■of fibre cells. Thus Ohara (1933b) finds a beaded appearance in silk 534 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III similar to that found in cellulose fibres, when the central portion, expanding powerfully, is pressed through weakened spots of the skin layer (Fig. 169). Furthermore, the central zone can be split up bv bromine lye (hypobromite) into fibrillae, which then disintegrate into short bundles of fibrillae, as in cellulose fibres. These facts are im- portant, in that they imply, contrary to Ludtke's statement (1936) about cellulose fibres, that beading and hydrolytic disintegration perpendicular to the fibre axis do not depend upon any partition across the fibre ; for there can be no question of the formation of any such hypothetical segmentation during the generation of the silk thread. Just as with other fibrous structures, a system of submicroscopic rodlets is to be expected; and, in fact, Ohara (1933a) found rodlet double refraction in silk. However, as the minimum of the curves he has published (n ^ i-47) does not agree with the average refractive index of silk fibroin, his measurements do not seem altogether reliable. Hegetschweiler (1948) finds (n,,)^ = 1.5960 and (nj^ = 1.5454 yielding the double refraction An = 0.0506 for silk. These figures give an average of (n^, -f 2nJ/3 = 1-565, which is quite in- compatible with n = 1.47 mentioned above. For this reason, Heget- schweiler (1950) repeated the imbibition experiments of Ohara and found that liquids which do not swell silk fibroin cannot penetrate and, therefore, do not change its birefringence. Since the cross- section of the fibroin thread is triangular (Fig. 168 a, p. 331), so that the thickness corresponding to a retardation of light F observed can only be measured after rotation of the thread through 90°, and as the swelling in aqueous solutions is considerable, it is very difficult to obtain reliable figures reflecting small changes of the birefringence of silk by the formula Zl n = F/d. If all the necessary precautions are taken and numerous measurements made in the same liquid in order to obtain reliable average values, it can be proved that silk fibroin is not a mixed Wiener body. This optical finding is borne out by electron microscope investi- gations (Hegetschweiler, 1950). Unlike native cellulose, silk fibroin does not consist of individual microfibrils. There is a distinct fibrillar texture, but the diameter of the visible strands depends on the method of preparation. The same fibroin threads show rather coarse (o.i ju) or very fine (0.0 1 /u) strands, or both types together with intermediate grades, depending on the way in which they have been hit during 2 PROTEINS 355 disintegration in the blender. There is a similarity here with rayon in which different types of fibrillar strands are visible in the electron microscope (Frey- Wyssling and Muhlethaler, i 949 c) without any evidance of individual microfibrils. It seems that during the spinning process a less regular, more compact body is formed than during o-rowth, when innumerable uniform microfibrils originate from a living matrix. There is therefore a pronounced difference between the submicroscopic texture of grown and spun fibres. Molecular structure. Silk fibroin consists of expanded polypeptide chains which crystallize in a chain lattice. This is why silk has a high tensile strength and a large intrinsic double refraction, similar to those of the chain lattice of cellulose. It is noteworthy that this similarity has no chemical background whatsoever, since silk fibroin and cellu- lose belong to quite different classes of chemical compounds. It is only the fundamental morphological principle of parallel macromolecular chains with a high polarizability parallel to the fibre axis which is responsible. This shows how important morphological considerations are for the analysis of the properties of high-polymer substances. According to Bergmann and Niemann (1957) silk fibroin consists of 2^ X 3^ = 2592 amino acids (mol.wt. --^-^ 220,000). Half of these are considered to be glycine, 1/4 alanine, 1/16 tyrosine, 1/2 16 arginine, 1/648 lysine and 1/2592 histidine. In addition to these constituents Drucker and Smith (1950) have found by paper chromatography small amounts of aspartic acid, glutamic acid, serine, threonine, valine, leucine, phenylalanine and proline. From viscosity measurements they assign a molecular weight of 33,000 to fibroin, which is almost one order smaller than that of Bergmann and Niemann (1937). This discrepancy is no doubt due to the fact that the determination of amino acids in very small quantities is open to considerable error. Since three quarters of silk fibroin consists of the smallest amino acids glycine and alanine, a relatively simple chain lattice can be derived from X-ray analysis (Meyer and Mark, 1928), if the other amino acids (tyrosine etc.) are considered to exist as amorphous substances without participating in the crystal lattice. This view is supported by the observation of Drucker and Smith (1950) that tryptic hydrolysis of short duration leaves glycine, alanine and serine, i.e. the simplest amino acids, undissolved, whereas all the other amino acids are found in the hydrolysis liquor. 356 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III In profile the crystallized chains present the familiar picture of the zig-zag line with consecutive CO, NH and CH groups (Fig. 170a). AsTBURY (1935a) calls the distance between neighbouring chains the "backbone spacing" and it measures 4.5 A, It may therefore be said that the depth which a chain requires in the plane of the zig-zag line, \ / RCH NH / NH \ CO / CO \ 'CHR NH RCH \ / NH \ CO Profile CO RCH / NH \o- / RCH \ NH / CO RCH / NH \ CO- / I -co- I CH I °^ NH ^ :;r CH, ■CO- I -CH I NH C>5 ,„J CO CH-CH,—(~^0H I NH I —-co I to > P Front a) b) Fig. 170. Molecular structure of silk fibroin. or the "backbone thickness", amounts to 4.5 A. The side chains of the CH groups are not seen in the profile view of Fig. 170 a, as they stand off, like ribs, perpendicular to the back- bone plane, suggestive of a vertebrate skeleton. The thread molecule has, therefore, to be seen from the front to get the side chains in their proper place (Fig. 170b). Thus the kinked chain appears as a straight line with foreshortened valence bonds, while the side chains lie in the plane of the drawing. The glycine residue can scarcely be said to have a side chain, which in this case is represented only by the insigni- ficant H atom; but with the alanine residue it consists ofa methyl group. The side chains are not fitted in pairs like actual ribs but point alter- nately to left and right on consecutive CH groups, with the result that every two neighbouring amino acid residues together form a morpho- logical unit, which in the X-ray diagram becomes the fibre period. In silk fibroinitamountsto6.95 A. From this it may be concluded that the length of each backbone segment, or in other words the extension of each individual amino acid member, is 3.5 A. This length is quite irrespective of the nature of the amino acids in the primary valence chain. Thus all the entirely different components, glycine, alanine and even tyrosine, represent, as members of the chain, sections of the thread molecule of exactly the same length (Astbury, 1933b). They can therefore be interchanged without thereby causing any alteration in the fibre period or the backbone thickness. The distance apart of the main chains depends on the length of the side chains which, as may be seen in Fig. 170 b where tyrosine is 2 PROTEINS 337 added, is very unequal. In order to preserve rigid regularity in this respect as well, Meyer and Mark (1950) assumed, as mentioned above, that only glycine and alanine residues form crystallized silk iibroin. It does not seem likely that this view can be maintained, for up to the present it has not been possible to define an undoubted elementary cell of the crystal lattice as in cellulose (Kratky and KuRiYAMA, 1951; Sakurada and Hutino, 1935; Brill, 1943). The reason may be a certain irregularity caused by the side chains of other amino acids. Friedrich-Freksa, Kratky and Sekora (1944) treated silk fibroin with iodine and found by X-ray analysis a new period of 70 A perpendicular to the fibre axis. As it is likely that the iodine is in- troduced into the tyrosine residue, every 20th amino acid of the poly- peptidic chain should be tyrosine. This would agree with the statement of Bergmann and Niemann (1937) that out of 16 amino acid residues one is tyrosine. It is therefore probable that tyrosine belongs to the crystallizing polypeptide chains. The primary valence chains are held together by hydrogen bonds (see p. 148) forming a chain lattice (Brill, 1941). In the glands of the silkworm the fibroin exists probably as globular protein called fibroinogen. Kratky, Schauenstein and Sekora (1950) find that air-dried glands yield an X-ray diagram similar to F-actin (see p. 352). It is called silk I, whereas the usual diagram is silk II. By stretching, silk I can be transformed into silk II of the spun thread. Only the transition silk I -> silk II has been observed, the reverse being apparently impossible. This favours the view that the formation of the silk thread consists in the denaturation of an originally soluble globular protein. If these protein molecules contain a certain pro- portion of all the amino acids found in silk fibroin, it would be likely that the polypeptide chains formed by denaturation comprise not only glycine and alanine, but also the other amino acids. The portions of the chains with unwieldy side branches would then not crystallize (Fig. 54b, p. 70) and might therefore be more easily susceptible to hydrolysis than the smooth glycine and alanine portions of the chain which can crystallize. Mercer (195 i) finds that microfibrils of fibroin (100 A thick and 3 500 A long) are formed spontaneously from a solution of fibroinogen in water. This seems to be a favourable object for studying the trans- formation of a globular protein into fibrils with the electron microscope. 558 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES Ul c. Ho} ny Substances (Keratin) Microscopic structure and birefringence of hair. The great technicaf importance and the remarkable elastic behaviour of wool and other hairs were the incentive to research on keratin. Microscopically, hairs consist of three layers, viz., a scaly and un- pigmented epidermis which is covered by a very thin cuticle or epidermicula (Lindberg, Philip and Gralen, 1948; Schuringa and Algera, 1950), a thick, fibrous cortical layer containing pigment, and a parenchymatous pith. Sometimes there is no pith, as in Merino wool. The surface skin, which covers the cortex with scales that are ring- shaped or like roofing tiles, may likewise disappear owing to mechanic- al chafing, and yet the elastic and optical properties of the hair will not radically change. Their source is, therefore, the keratin fibre cells of the cortex, which consist of numerous tonofibriUae orientated in parallel. In the electron microscope the fibrillae can be seen to unravel into still finer subfibrillae (Reumuth, 1942). The tonofibriUae vary in length between 50 f.i and a few millimetres, being about 80 pL (Hohnel,. 1887) in sheep. They are usually flattened. Although a hair appears to be optically homogeneous, it is not comparable cytologically to- a single bast fibre, but to multicellular strands of bast fibres consisting of relatively short fibre cells, as they occur in Monocotyledons (sisal^ Manila hemp, etc.). . Unlike cellulose fibres, horn fibres are extremely elastic. In cold water a hair can be stretched reversibly by 50 to 70^0, whereas bastr fibres of good fibrous structure break when stretched only a feu^ per cent. The elastic elongation of the hairs is especially impressive under the polarizing microscope (Pochettino, 1913). Although the cross-sectional area of the hair decreases owing to the elongation, the retardation increases considerably, and this is apparent from the sharp- rise in interference colours. It is a fascinating spectacle to watch the polarization colours of weakly pigmented (fair) hair changing as the hair is stretched and released. Whereas photo-elastic effects of this kind, however, are usually brought about by slight changes of distance in the crystal lattice which are not detectable by X-rays- (Wiener, 1926b), the molecular frame of keratin is completely re- formed during elongation. In curly wool the stretched outer side takes basic dj-es (Janus green^ neutral violet, pyronine) more easily, and has a lower refractive index. PROTEINS 559 in a radial direction, than the inner side of the curl (Ohara, 1958^ 1939)- Alol^cular structure of keratin. Astbury (1935 c) has demonstrated that stretched hairs produce quite a different X-ray diagram from unstretched ones. The difference is especially evident when the \ NH CO' rf CO-' \ CHR CHR NH ; II I5 NH CO C0-- CHR NH fi- Keratin y CH.R)-C0-NH- NH I CO ■R':CH '■' I NH CO CH^ CH (R\-NH' CO-\R)CH ^ NH CO . I IR)CH NH I ^CH NH I CO CH® '. ..■■' CO ■,RyCO-NH-(§)CH'' a-Kerafin b) Fig. 171. a) ^-keratin (after Astbury, 1933c). h) Folding of the polypeptide chain; R side chains (after .'\stbury and Bell, 1941). elongation takes place in a vapour-saturated chamber at 100° C, where about 100% elongation can be attained. The X-ray picture shows the distance between the members of the chain to be 3.38 A. This tallies well with the chain period of silk fibroin, viz. 3.5 A and it may therefore be assumed that elongated primary valence chains of poly- peptide thread molecules are also present in stretched wool. As the fibre period in unstretched wool is 5 .06 A, some other modification, which Astbury designates as a-keratin, must be involved. The keratin in stretched wool is known as /5-keratin. By folding the polypeptide chain, he derives a-keratin from jS-keratin, arguing that by the mutual attraction of two NH and CO groups separated by five valence bonds, pseudo-diketopiperazine rings are liable to be formed. Taking into account the rules of distance, the fibre period of a-keratin for a chain 34© FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III thus folded comes to 5.06 A. As the diagram shows (see Fig. 171a), the chain length is doubled at full stretch (100% elongation). The ingenious theor)^ of folding to form piperazine rings is con- fronted with steric difficulties ; for the side chains R, which point in the same direction, come so close together that they hinder each other spatially. Astbury and Bell (1941) have therefore drawn up a new folding diagram for the /5-a transformation, which satisfies the following conditions : I. The a-form must be about half as long as the ^^-form. 2. The density must remain practically constant. 3. The folds must repeat at a distance of about 5.1 A. 4. The side chains must stand out alternately on one side and the other of the plane of the fold. 5 . The folds must be nowhere so sharp as to have insufficient room for the side chains. This diagram is reproduced in Fig. 171b. Side chains pointing upwards are marked R enclosed in a full-line circle and those pointing downwards by R within a dotted circle. The side chains standing on the same side form groups of three, which in the diagram appear as the angles of the triangles indicated. The R side chains are particularly important. If hairs stretched in steam at 100° C are dried in the extended state, the elongation loses its reversibility and is retained. The side chains of neighbouring poly- peptide chains enter into spatial relationship and connect the primary valence chains to a kind of grid (Fig. 172). The distance between the bars of the grid is 9.8 A; hence the side chains, which at intervals of 3.38 A stand off more or less perpendicularly from the primary chains to the right and left, should have half that length. The thickness of the grid corresponds to the backbone thickness of the stretched, zig-zag polypeptide chains and is therefore 4.5 A. Glutamic acid, arginine and cystine are among the most important products of the hydrolysis of wool (see Fig. 88, p. 133). Assuming amidic linking between glutamic acid and arginine, there will be a kind of rung linking two primary valence chains, as represented in Fig. 172 b. Retaining the tetrahedral angle, this side connection would be about 12.5 A long. It is, therefore, of the order of magnitude of the length found by X-ray measurement, viz., 9.8 A, for it is quite conceivable that the chains may somehow be shortened. Cystine is the most interesting of the three. This contains two amino acid residues united by a sulphur bridge. It is assumed (Astbury, PROTEINS 341 1933c; Mark and Philipp, 1957) that such sulphur bonds hold the polypeptide chains together in keratin, for sulphur plays a similar part in vulcanized rubber. It connects the free polyprene chains of the raw rubber laterally, in this way producing a molecular frame, and thus enhances the elastic properties of the raw rubber, while 1 1 ' 1 1 1 1 1 1 1 1 1 1 — Sjde__ chains CO "1" 6 — tt" ft8/J :i o) NH NH CH' I CO- I NH RCH ho I NH ^: CO I CH, CO I (^0 NH CHy CH2 I ^CHl'^.N'H'^f^^CtC '^"^ Glutamic I acid -^0 Arginine j^^ I CM. CH. S S' Cystine ho NH -ho CHR ho f RCH Fig. 172. a) Keratin frame as lattice grid; b) side chains of keratin. its plasticity deteriorates. If too many sulphur bridges are intro- duced, however, the material will lose its elasticity, being "vulcanized to death", and hard rubber or ebonite results. Now there is some analogy between raw rubber and vulcanized rubber, on the one hand, and actomyosin (free from sulphur, p. 3 5 2) and keratin (containing sul- phur), on the other. By way of comparison, therefore, the tonofibrillae have been termed "vulcanized' muscle fibres, which would explain the loss of contractility and their great strength. Despite illuminating comparisons such as these, which are very helpful to a qualitative interpretation, there remain serious quanti- tative obstacles to a complete understanding of the submicroscopic structure of keratin. Above all, the length of the cystine molecule does not agree with the X-ray evidence as to the length of the keratin side chains. As is apparent from Fig. 172 b, the sulphur bridge is by no means long enough to span the distance of 9.8 A from primary chain to primary chain. Hence the molecular frame cannot be as simple and 542 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III orderly as it is represented in Fig. 172 a; possibly, therefore, some other amino acid besides cystine — say glutamic acid — assists in bridging this great distance. There is, however, another reason why the molecular frame is unlikely to be a simple structure. Astbury (1933 c) advances plausible arguments to show that, in the re-transformation of the /5-keratin of stretched hairs into the folded a-keratin, side bridges must be broken off. This, with a planar molecular frame, would be avoidable only if all the parallel 2ig-2ag polypeptide chains could be folded simultane- ously perpendicular to the projection plane of Fig. 172a without breaking the cross links. If, however, the primary valence sheets are linked in various directions, the individual polypeptide chains can no longer be folded without breaking up the side-chain bonds. It is very significant that ordinary water is capable of disrupting the bonds in question in the case of /3-keratin; for a hair stretched to double its length and then dried has only to be placed in water to regain its reversible elasticity. This means that drying brings about only temporary, and not permanent, set. Nevertheless, if a hair elongated 100%, is left for half an hour in a steam bath, it loses the capacity to contract again to its original length, being now permanently set and retaining this imposed length even when wetted in what is known as "permanent set". This fact is put to use in the hairdressing profession, for it is only when the hairdresser succeeds in imparting permanent set to the /5-keratin produced at the curved places in the hair that he can claim to have provided a "permanent wave". The permanent setting of the /5-keratin is said to be achieved by the prolonged action of the steam, whereby so many strong bridges are laid between the keratin chains that hot water is subse- quently unable to disrupt them. Elod, Nowotny and Zahn (1940) oppose Astbury 's theory that keratin contains grid frames connected by sulphur bridges in the side chains. Treatment of the wool with metallic mercury will convert 50% of the keratin sulphur to HgS. Removing half of the -S-S- bridges should weaken the molecular frame, involving modification of the properties of the wool. This, however, is not the case and these investigators therefore assume that it is not the side chains which build up the frame but, as in silk fibroin, hydrogen bonds (see Fig. 98, p. 148) between the primary chains in the backbone planes (Nowotny a PROTEINS 345 and Zahn, 1942). The side chains, they say, stand perpendicular to the planes of the frame and it is therefore of no consequence if they differ in length. It is assumed that the grids form a laminar structure parallel to the surface of the hair or nail. Other arguments against the salt link theory are advanced by LiNDLEY (1950). The basic amino acid residues arginine, lysine and histidine have a constant ratio 12:4:1, whereas the other members of the keratin polypeptide (cystine, tyrosine, glutamic and aspartic acid) show considerable fluctuations depending on the wool sample chosen (Block, 1939). Fractional hydrolysis yields peptides of low molecular weight with numerous acid residues clustered together, whilst the basic amino acids are regularly distributed along the poly- peptide chains. MiDDLEBROOK (1951) thinks that cystine which amounts to about 1/8-1/4 of the total number of amino acid residues is concentrated in definite regions along the polypeptide chains, and that these regions, w^ith a periodicity of about 200 A, cannot assume the a-folds because of steric hindrance. Therefore, a-keratin would always contain peri- odical short segments with straight /^-constellation. If a hair, which has been stretched 100% and temporarily set, is placed, free, in a steam bath for a short time, it will contract, not only to its original length, but considerably further; a super con- traction takes place. This fact implies that the polypeptide chains in the a-keratin are not entirely free and independent of each other; rather, it would seem that they too are mutually stabilized by certain bridges. Apparently, however, the treatment breaks up these linkages and enables the polypeptide chains to fold far more than before. Restrictive lateral bridges of this kind are also supposed to be re- sponsible for the fact that a hair is only 50-70% extensible in cold water; they weaken in hot water and the polypeptide chains can then be fully stretched (about 100%). If keratin is exposed for a short time to the action of hot water or vapour, connecting bridges between protein chains are evidently broken down. Since dilute caustic soda similarly loosens the chains, this might be a case of hydrolytic decomposition of acid amide bridges. Yet the self-same treatment, if more prolonged, will facilitate the formation of new, stronger bonds. In view of the theory regarding the structure of cytoplasm developed in this book, this behaviour is 544 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III very significant, as it shows how readily the frame of proteins can be destroyed and built up again. Seeing that hot water suffices to initiate this process in keratin, it is not difficult to imagine how, in the far more labile cytoplasm, the protein thread molecules are constantly forming new combinations and side bonds, while others are con- tinually being broken down, so that a definite molecular framework is always in existence, despite the apparent liquid state of the material. Fine-structure of finger nails. Finger nails are built up of submicro- scopic fibrillae. X-ray analysis shows that the keratin fibril lae run, not parallel, but perpendicular to the longitudinal axis of the nail (Derk- SEN, Heringa, and Weidinger, 1957). As in the elongation growth of the plant cell wall, therefore, the micelles are orientated perpen- dicular to the direction of growth. The alignment of the micellar strands, therefore, is not a passive process due to the forces of growth pushing the nail forward; there are special formative forces at work, building up submicroscopic textures with due regard to their future functions. By maceration with NaoS a thin, 100 A thick membrane can be detached from the surface layers of human skin and finger nails. It is compared with the epicuticle of wool (Lagermalm, Philip and Lindberg, 195 i). Fsather keratin. Not all horny substances are naturally in the state of a-keratin. Instead of the fibre period for mammalian hair, viz., 5.06 A in the direction of the primary chain, that of quills in the un- extended state is 3.1 A (Astbury and Marwick, 1932). By elongation it can be increased continuously and reversibly to 3.3 A but, if subjected to further elongation, the quill breaks. From this fact it may be concluded that the polypeptide chains in quill keratin are stretched approximately in the same way as in elongated hairs or in silk fibroin. The fact that the length of the members of the primary chains is neither 3.38 A nor 3.5 A is said to be due to slight corrugation (so-called "primary folding") of the polypeptide chains in the feather keratin, owing to a certain interaction of the side chains. This slight primary folding is also supposed to be responsible for the shortness, as compared with silk fibroin, of the amino acid residues of ^-keratin. The far sharper kinks in the a-keratin chains are distinguished from this slight corrugation as "secondary folding". Thus the sclero- protein of quills is a modification of keratin in which there is no PROTEINS 545 secondary folding. The keratin primary valence chains are therefore used by the animal body for the building of the horny tissues, either heavily folded, or in more or less stretched condition. d. Connective Tissue {Collagen) Molecular structure of collagen and elastoidin. Tendons and decalcified bones consist of the gelatinous protein collagen. Glue and gelatin are relatively little changed decomposition products of this insoluble frame substance which have become soluble in hot water owing to sHght hydrolytic degradation. TABLE XXX CHEMICAL COMPOSITION OF COLLAGEN (SCHAUENSTEIN AND STANKE, I951) 0/ /o Arginine . . 4-9 Histidine . . 0.5 Lysine . . . 2.8 Oxylysine . . 0.8 Glutamic acid 4.0 Aspartic acid 2.6 Glycine. . . 34.6 Alanine. . . 10.5 ; Valine . . . Leucine. . . Proline . . . Oxyproline . Phenylalanine Serine . . . Threonine . /o 2.8 4-1 12.7 II. 2 2.4 5.0 98.6 Collagen is a protein the chemical composition of which differs remarkably from the amino acid content of reserve proteins. It contains a considerable amount of proline and oxyproHne (Table XXX) but no tyrosine, tryptophane, cysteine nor methionine. Tendons and elongated gelatin both produce the same X-ray pattern (Gerngross and Katz, 1926). It shows 8.4 A as the fibre period which, divided among three amino acid residues, shows the length of the members of the primary chain to be 2.8 A. Moreover, there are two interferences on the equator of the diagram, which correspond to 4.65 A (backbone thickness of the primary chain) and 10.0 A (length of the side chains). The resemblance to the conditions in /5-keratin is striking, except that, as compared with the amino acid residues of silk fibroin and of keratin, the primary chain period of 2.8 A would appear to be rather short. This may be due to the 346 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III N / CH COH- I N II -COH N I HC I ■CH-. \ '\ / CH, CH. / CH, co- COH- CHR \ / CH \ Profile COH- HOCH \ CO I CH^ N II - COH I CHR I N II COH /"^^Ih Proline Glycine ■ Amino acid R presence in the collagen of about 24% of proline and oxyproline in addition to 34% of glycine. The many five-membered rings cannot, of course, all act as chain end groups (see Fig. 88, p. 133); they must surely be built into the primary chains (Fig. 173), causing con- siderable primary folding (Astbury, 1940). Collagen, like the other frame proteins, appears after all to be built up according to the diagram of polypeptide chains of indeter- minate length. Tautomeric rearrangements help to explain the shortening of the members, for if within the stretched chain the hydrogen of every second NH group is transferred to the neighbouring CO group, double bonds -N = C(OH)- are formed which entail the stereoisomeric poss- ibilities of the cis and trans configur- ation. If the cis position is assumed, the members of the chain are short- ened to the value of 2.86 A ascert- ained experimentally (cf. Halle, 1937; Ch ampetier and Faure - Fre- MiET, 1938). The enolic peptenol form [^ C (OH)] of the polypeptide chain (Fig. 173) has been shown by UV absorption, since ^ C = C<^ bonds yield a characteristic UV band between 2400 and 2600 A. Since the aromatic amino acids tyrosine and tryptophane, which have over- lapping absorption bands, are absent, the peptenol group can be studied in collagen by this method better than in any other protein (ScHAUENSTEiN and Stanke, 195 i). There is a similar small fibre period of 2.9 A in elastoidin (Cham- petier and Faure-Fremiet, 1937), which is the frame substance of the fin rays oiSelachii (Faure-Fremiet, 1936). Its thermal and sweUing properties are comparable to those of collagen, from which elastoidin is distinguished by slight chemical differences in resistivity to trypsin and by sulphur content. Optical and swelling behaviour of tendons. Optically, tendons and CH:; I ,N I Oxyproline Front view Fig. 173. Diagram of a gelatin chain. 2 PROTEINS 347 welatin filaments are positively uniaxial as referred to the fibre direction, i.e., the same as silk and hairs. Rodlet birefringence is also evident if the tendons are tanned before imbibition (Kuntzel, 1929). The tendons are very liable to swell in the presence of most imbibition liquids, or to shrink (e.g., with xylene). Collagen behaves peculiarly on tanning; for whereas the optical character of the tendons is retained with mineral tanning materials (chromic salts) and formol, it is reversed and becomes negative with pyrogallic tanning agents (tannin, sumach) and other phenols (trinitrophenol) and aldehydes (eugenol, cinnamic aldehyde). Schmidt (1934) imagines that the optical negative reaction is brought about by orientated adsorption, as the non-tanning univalent phenols and aldehydes may be washed out again, whereupon the normal optically positive reaction returns. Personally, I am inclined to believe that it is rather a matter of chemi- cal- changes in the side chains. Tanning depends upon the permanent connecting of one polypeptide chain molecule to another by strong side-group linkages. Moreover, the pyrogallic tanning agents must thereby change the polarity of the side groups in a manner similar to what takes place in the nitration or acetylation of cellulose. In view of the lability of many side chain reactions of the polypeptide chains, it is not surprising that washing out of the non-tanning phenols should easily upset the chemical changes brought about by trinitrophenol, eugenol, etc. Rodlet birefringence and X-ray analysis thus provide evidence for the submicroscopic fibrous structure of tendons. It is not only the strange optical behaviour of tendons which has for long attracted attention (v. Ebxer, 1894), but also their re- markable swelling power. In water they swell by 50% in thickness, which, as X-ray evidence shows, involves expansion up to 35 % of the crystal lattice (Kuntzel and Prakke, 1933), while the fibre period remains unchanged. Hence the swelling is not intermicellar as in cellulose, but intramicellar, inasmusch as the individual primary valence chains are pushed apart. This explains why the swelling of tendons may assume unprecedented dimensions. In dilute acids and alkalies, which obviously completely hydrolyze the side-chain bonds, they are liable to swell 550*^0 in thickness, though admittedly they shorten at the same time by 30%. Despite this shortening, the increase in volume due to the infiltration of fluid may amount to as much as 4500% (Kuntzel and Prakke, 1933). 348 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III Reduction in length becomes more striking when the tendons are placed in hot water (60 to 70° C). They suddenly contract, while swelling, and at the same moment the birefringence and X-ray dia- gram vanish. This unusual reduction in length imparts rubber-like elasticity to the tendon. After careful elongation the X-ray diagram reappears and continued stretching will finally restore and establish the inelastic collagen fibre. All this resembles the behaviour of rubber which, when unstretched, produces no X-ray diagram, but gives a pattern after it has been considerably stretched. Meyer and Mark (1930) point out another interesting property common to both materials. If the contracted tendons or unstretched rubber be frozen in liquid air and the objects be then smashed, they crumble to a friable mass, like sand; whereas under similar treatment native tendons or elongated rubber will split up into a fibrillar mass. From this it may be inferred that the polypeptide chains of the collagen fibres contract, as in /9-keratin, and fold up. But whereas folding of the ^ -> a-keratin type is limited, with collagen it is so violent that the straight protein chains shrivel up completely. Evidently, the impulse of polypeptide chains to shorten in the free state is very widespread and, if means are found to make this process reversible and to regulate it, a model will be provided for the contractile muscle fibres. Suhfnicroscopic striation of collagen fibres. While collagen fibrils are perfectly smooth in the ordinary microscope, they appear to be striated in the electron microscope. This striation was first reported by ScHMiTT, Hall and Jakus (1942) and by Wolpers (1944). The period of the cross-bands in collagen fibrils of the human tendons and human skin is 640 A (Schmitt, Hall and Jakus, 1943; Gross, 1950). This corresponds to the macroperiod found in kangaroo tendons by X-ray small angle diffraction (Bear, 1944). X-ray analysis has also revealed the remarkable fact that the extended collagen chains do not form a three-dimensional lattice, the direction of their side chains changing arbitrarily in the chain lattice (Bolduan and Bear, 1950). Pratt and Wyckoff (1950) have shown that in the particularly clean fibrils of collagen from dog heart, the segments are bordered by pairs of cross striae (Fig. 174). Sometimes a third cross-band is seen in each segment between the pairs. Then the fibril appears continuously cross-striated with a period of ca. 640 A/3 = 210 A. PROTEINS 349 These authors think that the third cross-band is due to remains of a second transverse system of fibrils which bind the separate collagen fibres into a fabric-like system. It is likely that it is the pairs of cross striae which adsorbs more easily silver than the rest of the segment (Dettmer, Neckel and Ruska, 195 i). 0^ ^'. - Fig. 174. Striation of collagen fibrils, 34,000: i (from Pratt and Wyckoff, 1950). The reason why collagen fibrils display a submicroscopic seg- mentation is obscure. Possibly there is some relation to the globular state of proteins in solution. Bahr (1950) and Noda and Wyckoff (195 1) succeeded in reconstituting tendons dissolved in dilute acetic acid into collagen fibrils by precipitation of a dilute collagen solution of 0.75 to 0.05% with salts, (0.7 to 1.5% at p^ 3.8 to 7.0). This reconstitution furnishes segmented collagen fibrils with a period of 635 A or 650 A which can be dissolved again by dialysis against water and acetic acid. Since dissolved collagen represents a globular protein, as known from gelatin (see p. 93), precipitation may join these spheres in a linear way causing beaded chains ; if such chains associate laterally to form fibrils, denser and less dense cross-bands are likely to be produced. In the work of Schmitt, Hall and Jakus (1942, 1943, 1945) the dense segments are marked A, the more transparent, B. Upon arti- ficial elongation the B segments increase in length at the expense of the A segments ; the period can be raised to as much as 6000 A. It is :Supposed that the polypeptide chains are more tightly folded in the 550 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III A segments than in the B segments and that they unfold partially when stretched. When collagen fibrils are stained with phosphotungstic acid, the electron microscope reveals not merely one dark and one bright segment per period, but a series of bands (e.g., 5) within the dark segment. Apparently these bands combine preferentially with the phosphotungstic acid. Suhmicroscopk structure of hones and teeth. Bone represents a complex system of collagen fibres, its inter- fibrillar substance being calcified by hydroxyl-apatite Caio(P04)6(OH)2 (Brandenberger and Schinz, 1945). It is a Wiener body, which can only be imbibed, however, with liquids of variable refractive indices, if one of the two com- ponents is destroyed. This can be done either by decalcification or by ignition of the organic com- ponent. AscENZi (1950) has mea- sured the form birefringence of the organic substance (ossein) and the inorganic substance. The re- sult is represented in Fig. 175. The ossein has almost no intrinsic bire- fringence, whilst that of the incin- erated bone is strongly negative. This is due to the optical properties of the hydroxyl-apatite which cry- stallizes in optically negative hex- agonal prisms. The minimum of the rodlet birefringence curve at n = 1.600 is caused by the optical properties of apatite (1.634 — 1.638 = — 0.004). In the electron microscope the collagen fibrils are visible with their striation (Rutishauser, Huber, Kellenberger, Majno and Rouil- LER, 1950; Huber and Rouiller, 195 i). Teeth have a similar submicroscopic structure to that of bones. Of special interest is the enamel which covers the dentine as a specially 0.009 0.008 0.007 0.006 0.005 - O.OOi, 0.003 0.002 0.001 0 001 0002 ■ 0.003 -0.00^ 0 0 0 0 0 0 0 0 0 10 0 IT) 0 10 0 LO 0 U-) IT) U1 0 10 x-~* r-^ Fig. 175. Birefringence of human femoral diaphysis (from Ascenzi, 1950). A) In- organic bone fraction, B) ossein, C) total bone. Abscissa: refractive index of the im- bibition liquid. Ordinate: birefringence. Z PROTEINS 351 hard and resistant layer. It consists of parallel prisms of hydroxyl- apatite orientated perpendicular to the enamel surface. Its optics have been investigated by W. J. Schmidt (1923, 1936/37). In a very young state this layer is optically positive with respect to the axis of the prisms, whereas in full-grown teeth the enamel assumes an optically negative character. This change is explained by the following facts: In the embryonic state the enamel prisms represent a Wiener body with submicroscopic spaces between submicroscopic crystallites of hydroxyl-apatite; hence, its positive double refraction is caused by 'marked rodlet birefringence. Later, when the enamel hardens, the submicroscopic spaces are filled with material of a refractive power similar to that of the crystallites ; consequently the form birefringence disappears and the optically negative intrinsic birefringence of the hydroxyl-apatite becomes visible. This behaviour proves that the optical axis of the submicroscopic crystallites must run parallel to the axis of the microscopically visible prisms. It had been assumed that the filling material would be exclusively inorganic, but the electron microscope has shown that it is not. Scott and Wyckoff (1946/47) have developed a method for preparing thin replicas of pofished and slightly etched tooth sections. In order to obtain undamaged replicas they must be freed by dissolving the section in 18 % HCl and 2 % pepsin, whereupon the replica is shadowed. In such preparations the microscopic enamel prisms appear to be surrounded by a thin organic sheath and inside the prisms there is a very fine organic matrix (Frank, 1950). Enamel is not therefore simply an inorganic coat of the tooth, but contains an organic frame as well. This explains why even completely intact teeth are subject to decay. Elastic tissue. Elastic tissue as found in the back of the neck (liga- mentum nuchae) of the vertebrates differs from the connective tissue of collagen in several characteristic properties. Being highly elastic and resistant to tryptic digestion, its protein has been given a special name, viz., "elastin". Threads of elastin have little birefringence, but their double refraction can be enhanced by stretching (Schmidt, 1924). It was therefore thought that elastin w^ould behave like rubber, with dis- ordered chain molecules in the relaxed state and parallel molecules in the stretched state (cf. Gross, 1949). 552 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III €. Muscle Fibres (Actomyosin) Proteins of muscle fibres. Fresh striated muscle contains about 20 % of protein. On extraction of minced muscle with water, the soluble protein myogen is dissolved; but this protein does not appear to be involved in contractility. The contractile substance is contained in the insoluble fraction. If treated with a slightly alkaline salt solution the protein myosin can be extracted. Its I.E. P. is 5.3. Szent-Gyorgyi (1943) succeeded in obtaining crystallized myosin and established in this way that ordinary myosin solutions are contaminated with another protein, actin. Under the electron microscope crystallized myosin appears to be a fibrillar protein (Astbury, 1947/49). Its molecular weight is 1.5 million (Snellman and Erdos, 1948). When myosin is properly extracted from muscle tissue, the main part of the acfin remains in the residue. After drying this solid fraction with acetone, the actin can be dissolved (I.E. P. 4.7). The solution is perfectly clear and has a low viscosity. When left in the presence of KCl, it becomes more viscous and ultimately turns into a thick thixo- tropic gel. This gelation is due to a transformation of globular protein. Both modifications are visible in the electron microscope. The globular actin has been called G-actin and the fibrillar modification F-actin. The particles of G-actin are ellipsoidal with the dimensions 300 A X 100 A. Rosza, Szent-Gyorgyi and Wyckoff (1949) have shown how these particles form F-actin in situ by linear aggregation. The filaments of F-actin are 100 A thick and appear to be segmented with a period of 300 A. They aggregate laterally forming cross- striated bands. The G-actin as seen in the electron microscope would have a molecular weight of about 1.5 million, whereas, according to measure- ments in the ultracentrifuge, it consists of only 4 Svedberg units (MW = 70,000; Straub, 1948). The particles of G-actin visible in the electron microscope therefore consist of about 24 actin molecules. Neither myosin nor F-actin is contractile. But if these two proteins are brought together they react with each other forming the con- tractile substance F-actomyosin. There is an optimal reaction with a ratio of 2.5 parts myosin to i part F-actin. Snellman and Erdos (1949) conclude from this fact that there is a stoichiometrical ratio of these two components in the contractile muscle protein. 2 PROTEINS 355 When adenosine triphosphate (ATP) is added, F-actomyosin con- tracts violently. Under the electron microscope F-actomyosin consists of fine filaments and, after treatment with ATP, coarse threads. However, as there is no change in the X-ray pattern, the syneresis which occurs has been declared to be intermolecular and not intra- molecular (Perry, Reed, Astbury and Spark, 1948). A gel of 2-3% actomyosin throws out so much water by dehydration as to become a dense gel of 50% protein. The mechanism of this contraction is not vet thoroughly understood. Optics of striated muscle fibres. The safest way to assess the micro- scopic structure of the highly differentiated striated muscle fibre is between crossed nicols (Vles, 191 i; v. Muralt, 1933; Schmidt, 1937a). This circumvents many sources of error, such as the com- plicated diffraction phenomena of striated sj^stems (Pfeiffer, 1942 b; L.\NGELAAN, 1 946), and the changes in structure which are greatly, though sometimes wrongly, feared in the fixation of tissues. The muscle fibre is lo to loo ^ in width and is enclosed in a thin skin, the sarcolemma. It disintegrates into optically resolvable fibrillae about I [x thick and at roughly 0.5 ji distance from each other. The visible fibrillae consist of bundles of parallel submicroscopic ele- mentary fibrils (HiJRTHLE, 1931). The sarcoplasm, which surrounds the fibrils on all sides, lies in between the myofibrils. Essentially it consists of muscle albumin, or myogen, while the fibrils are identical, in the main, with muscle globulin, or actomyosin. The sarcoplasm is always isotropic, but myofibrils are birefringent and exhibit the familiar segmentation into bright, so-called Q and A bands and dark, very weakly birefringent (usually called isotropic) I bands, which are subdivided by a stronger birefringent Z band. There are accumu- lations of nucleic acids of the adenyl nucleotide type in the semi- isotropic I sections (Caspersson and Thorell, 1941). The re- markable part of this structure is that all the fibrils of a muscle fibre, though independent, have their bright and dark bands at exactly the same level, with the result that the entire fibre is evenly striated. The coincidence of the strongly and weakly birefringent bands is due to the division of the individual fibrils after the striation of the original mother fibrils has occurred. Despite the conspicuous optical differentiation, the fibrils are not transversally subdivided, but run in uninterrupted succession through the entire length of the fibre. Their 354 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES HI cleavability, which betrays no mechanical inhomogeneity at the boundaries of the segment, is an argument in favour of their uni- formity. Further, very young fibrils are uniformly birefringent (later the striation appears gradually differentiated from the middle towards the extremities), while the cross striation may disappear in explanted skeletal muscle cells through dedifferentiation (Schmidt, 1937a, pp. 215, 223). Whereas the fibrils are probably continuous, the sarcoplasm appeaxs to be subdivided by transverse septa; for in the centre of the dark I band there is always a narrow Z band, easily identified by staining, which shines brightly between crossed nicols (Fig. 176). It is supposed to be a cross membrane, continuous with the sarcolemma, the myo- fibrils thrusting through it without hindrance. When the muscle contracts, these regions do not thicken appreciably, so that the sarcolemma is thrown into festoons. The segment of the myofibril from one Z band to the next is called the sarcomere. Its length is about 2 //. In a growing muscle fibre, the sarcomeres are added to the end of the fibre originating from one single cell. The sarcomeres at the two ends are less differentiated during growth than in the middle of the fibre (Haas, 1950). On both sides of the Z band slightly birefringent N bands occur, often joining the Z band. Matoltsy and Gerendas (1947) suppose the lack of optical anisotropy in the I band to be caused by the inter- calation, between the myofibrils, of an optically negative substance,, called N-substance, which compensates the positive double refraction of the actomyosin (Gerendas and Matoltsy, 1947). The UV ab- sorption of the N-substance is the same as that of nucleic acid, which is an optically negative substance (see p. 220). Muscle fibres extracted with 0.3 Af KCl, which dissolves myosin, lose their isotropic bands (Snellman and Gelotte, 1950). The retardation of the Q bands in the fibre decreases considerably during contraction, notwithstanding the appreciable increase in thickness; the optical term for this is negative fluctuation. The fact established by v. Muralt (1932) that negative fluctuation also occurs with isomeric contraction — i.e., when the muscle is forcibly held to its original length during contraction — is of great importance. Besides intrinsic birefringence, which is manifested as birefringence of flow in myosin solution (v. Muralt and Edsall, 1930), the myo- PROTEINS 555 fibrils exhibit distinct rodlet birefringence (Stubel, 1925). It follows from this that the fibrils are not uniform in structure, but are of the class of rodlet composite bodies. Boehm and Weber (1932) produced composite bodies of this kind artificially by injecting myosin solutions into water. The resulting filaments displayed the same optical proper- ties, both qualitatively and quantitatively, as the Q sections of the myofibrils (Weber, 1934). It is surprising to find how well the measured birefringence agrees with that calculated from Wiener's formula (see p. 84), for the assumptions of Wiener's composite bodies are hardly applicable to hydrophilic micellar systems. Above all, the theory requires that there should be a well-defined phase boundary between the rods and the imbibition liquid, which there cannot be with a swellable protein which adds on water molecules to its macromolecular chains. A further assumption, which is more to the point in this case, is that the submicroscopic rodlets have practically unlimited length. Weber, it is true, assumes a particle length of only 500 A and Wor- SCHITZ (1935) has X-ray evidence for lengths up to 2050 A, but no reliance can be placed on X-ray determinations of particle length with dimensions beyond 500 A (see Frey-Wyssling 1937a, p. 376). It may therefore be assumed with equal justice that the optically identified rodlets are bundles of primary valence chains of unknown length which run parallel through the myofibrillae. X-ray analysis. X-ray analysis gives us some in- formation about the inner structure of elementary fibrils. Myosin filaments produce the same X-ray diagram as relaxed muscles (Boehm and Weber, 1932), which proves the identity between the fibrillar substance and myosin. Model experiments can therefore be carried out with m^^osin films and it is in this way that Astbury and Dickinson (1935 a) found that the X-ray picture of muscle protein corresponds to that of keratin. The a ?^ /3-keratin conversion can be attained by elongation, but in the relaxed muscle it is not the stretched Fig. 176. Striated muscle fibres between crossed nicols (by courtesy of Prof. W. J. Schmidt, Giessen). Wide Q sections and narrow, weakly luminous Z stripes. 356 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III jS-form, but the folded a condition that is found. True, the modifi- cation of myosin to the ^-form can also be forced upon the muscle by artificial extension (Astbury and Dickinson, 1935 b), but the a-form always occurs in the natural state. Hence it must be assumed that the polypeptide molecules in the relaxed muscle run, as in un- stretched hairs, in folded chains parallel to the fibre through the fibrillae. This is where the X-ray method is at a distinct disadvantage as compared with polarization optics, for it fails to distinguish the more strongly birefringent Q sections of striped muscles from the almost isotropic I bands. Fig. 177. Electron micrographs of striated muscle fibres (from Hall, Jakus and Schmitt 1946). Above: relaxed; below: contracted. 2 PROTEINS 557 Electron microscopy. The electron microscope provides a means of checkine the conclusions derived from the results of indirect methods. WoLPERS (1944) and Hall, Jakus and Schmitt (1946) find the following micrographs of striated muscle fibres (Fig. 177): the Q segment is dark and interrupted by a cross-band M, whilst the I segment is clear; i.e., there is a denser packing of protein in Q and a much looser arrangement in I. The most surprising result is the complete blackness of the Z zone in contrast to the lack of electron scattering in the adjacent N zones (Fig. 177 above). In that part we must assume the presence of heavy atoms and, as Caspersson and Thorell (1941) have found more nucleic acids in the semi-isotropic sections of the fibres, it is likely that phosphorus, besides metallic cations like potassium, is accumulated in the Z zone. It might also be possible that the Z zone has a special adsorbing power for heavy metals, since osmium fixation (Wolpers, 1944) or phosphotungstic staining of the fibres has been used in the previous treatment. In contracted muscle fibres much electron scattering material is found in the I band (Fig. 177 below). F. O. Schmitt (1950a) assumes that on irritation there is a migration of Q-substance into the I band, causing the much discussed reversal of striation. The microscopic myofibrils consist of parallel submicroscopic microfibrils of 100-150 A diameter. Like the myofibrils, these micro.-, fibrils run straight through the segments and across their border lines. For this reason earlier attempts to explain the weak optical anisotropy of the I bands by a disorientation of submicroscopic elements must be discarded. The microfibrils produce the X-ray interferences of both actin and myosin ; hence they are considered to consist of actomyosin (AsTBURY, 1947/49). X-ray diffraction discloses a long-range axial period of 400 A and a short-range spacing of 27 A, while in F-actin 54 A has been found (Schmitt, 1950a). In palladium shadowed electron micrographs Rozsa, Szent- Gyorgyi and Wyckoff (1950) offer evidence of the incrusting materials in the myofibrils. They find a heavy incrustation in the Z and M zones. Further, the whole Q band is rich in interfibrillar substance except two narrow zones, called H zones, adjacent to the M stripe which intersects the Q segment. Unexpectedly the I bands, with the exception of the Z stripe, are free from such substances. The authors consider the microfibrils to represent pure F-actin and discuss 358 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III whether myosin could be a part of the incrusting material. This does not seem likely, since the ratio of myosin to actin is 2.5 (or even 3) to I (Snellman and Erdos, 1949) so that myosin cannot be an ac- cessory substance in the muscle fibre, but must be incorporated in the fibrillar material. It is likely that potassium ions are part of the dense substance of the Q bands, which is rich in ash, as disclosed by microincineration. All incrusting substances can be removed by washing without disturbing the course of the microfibrils, whereupon a perfectly smooth myofibril results. Present information on the fine-structure of myofibrils is detailed and extensive, but still confusing. Matoltsy and Gerendas (1947) claim to have found an optically negative N-substance incrusting the I segments, whereas this segment is free from interfibrillar material according to Rgzsa, Szent-Gyorgyi and Wyckoff (1950), so that its semi-isotropy is difficult to understand. Further, on the ground of the negative fluctuation of the birefringence during contraction, it is generally accepted that the Q segments shorten more than the I segments. Hall, Jakus and Schmitt (1946), on the contrary, have observed in the electron microscope that the Q band of contracted myofibrils does not change, whereas the I band is shortened con- siderably, accounting for almost the whole contraction, which amounts to 40% of a sarcomere (relaxed about 2 /z, contracted 1.2 fj). By staining with phosphotungstic acid. Hall, Jakus and Schmitt (1945) were able to detect a submicroscopic banding in smooth muscle, the fibre period being 725 A. It would therefore seem that the banding of protein fibrils is a common property, resulting, as the electron microscope discloses, from the periodic dense and loose packing of protein or phosphorous substances. The mechanism of muscular contraction. There are several ways of at- tacking the important problem of muscular contraction: thermo- dynamic, chemical and morphological views may help to find a consistent explanation. The thermodynamic approach has tried to make the disorientation of molecular elements responsible for the liberation of energy when the fibre contracts (cf. Bailey, 1942). Bio- chemical investigations show, however, that the energy is liberated by the reaction of myosin and adenosine triphosphate, this nucleotide being dephosphorylated and the liberated phosphoric acid used for the phosphorolysis (see p. 314) of glycogen. The enzyme adenosine 2 PROTEINS 359 triphosphatase is intimately tied to myosin or may even be a part of this protein molecule (Needham, 1942 a, b; Potter, 1944). The intimate interrelation of the mechanics of the contractile muscle with chemical reactions is shown by Hill (1950). We have first to discuss the morphological side of the problem. When muscle contracts, the polypeptide chains coil up. Both extended and relaxed muscle have the a-keratin structure, which becomes dis- orientated on contraction (Huxley and Perutz, 195 i). Actually the same thing occurs as in the supercontraction of the keratin chains, with the difference, of course, that in the case of muscle the phe- nomenon is reversible and can be voluntarily induced. A relaxed muscle frozen in liquid air splits up into fibres, whereas a contracted muscle disintegrates into small lumps (Meyer and Mark, 1950). Furthermore, contraction wipes out the X-ray diagram. Roughly speaking, a contracted muscle is amorphous like unextended rubber, whereas in the relaxed state it is crystalline like elongated rubber. Notwithstanding the enormous mass of literature on the physio- logical processes involved in muscular contraction (Verzar, 1945; Faure-Fremiet, 1946), we do not yet (^-co^ know what special process it is that Hydration- Induccs the folded polypeptide chains to supercontract. K. H. Meyer (1929) ^fooj \<:Zi:>NH^ suggests that fundamentally it is the _,^„ >coo" mutual repulsion of groups bearing the — ^ ^^^^coo~ same charge, e.g.,-COO~groups, which M~ -^ V. prevents the chains in the relaxed muscle Ph^^ ^"^'^^ ^ from crumpling. This occurs when the Fig. 178. Contraction of the poly- chain is in repose at a Ph of 7.4, viz., in peptide chains in the isoelectric ^^ ^[y^^Unc medium (see Fig. I78). Now state (I.E. P.) (after K. H. Meyer, ^ , ic,2c,), if by some physiological process the pjj of the muscle serum is reduced to 4.7, which corresponds to the isoelectric point of the protein actin, the amino groups become positively charged and the groups with the opposite electric charge are attracted to the point of contact and the chains coil up. Kuhn and Hargitay (195 i) have calculated these attractive and repulsive forces foi the case of polyacrylic acid, which contracts in an acid and expands in an alkaline medium. They find 360 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III values compatible with the stress measured in contracting threads of polyacrylic acid when placed in o.oz N HCl. The matter is probably not quite as simple as this, for the charges of the polypeptide thread molecules are not in the primary chain, but at the extremity of the end groups of the side chains. In a later work, Meyer and Picken (1937) prove by thermoelastic investigations on stretched muscle fibres that, in a state of rest, the polypeptide chains are mobile as in a liquid, whereas fixed bonds are established as soon as the muscle is irritated ; thus the molecular framework of the muscles passes from an apparently "fluid" to a solid state. It should be noted that the comparison with rubber ceases to be valid under these circum- stances, for in that material the polyene chains are, conversely, more mobile in the contracted state and are interlocked in the elongated con- dition. The interesting reaction involved in muscular induration must surely take place between the end groups of neighbouring side chains. This is a good example demonstrating the consistency of our theory of junctions. Contracted muscle fibres exhibit an extreme gel structure, whereas relaxed fibres show a less tightened structure. We may thus compare muscle relaxation with the transformation of the plasmagel to the plasmasol in protoplasmic flow, when junctions must similarly be freed to allow displacement of the structural elements. f. Nerves (Neurokeratin and Neuronin) The wjelin sheath. Myelinated nerves in Vertebrates consist of a central strand enveloped in a highly birefringent sheath. The bire- fringence of this sheath is produced by the embedded myelin, which produces the myelin forms described on p. 54 upon the addition of water. Like myelin tubes, the mvelin sheath is optically negative as referred to the axial direction. Referred to the radial direction, however, the birefringence is positive. Thus in a cross-section through the nerves the sheath shows a positive cross, while the axoplasm appears as isotropic. Since myelin comprises lecithin (Fig. 47, p. 56), cephalin, cholesterol (Fig. 92, p. 138) and other anisodiametric optic- ally positive molecules, they must, judging by the birefringence, be orientated in the sheath with the longitudinal axis running radially. Isolated myelin substances produce X-ray periods corresponding to double the molecular length. There must therefore be bimolecular lipid layers in the nerves. The thicknesses of the layers are given in PROTEINS 361 Table XXXL The averay-e distance between the moIecul?r chains is 4.8 A (BOEHM, 1933). Small-angle X-ray diffraction furnishes layer periods of 186 A for fresh and 1 5 8 A for dried mammalian nerves. This shrinkage shows that hydration water lies between the lamellae. Since a drv double layer of neural myelin is only 66 A thick, it is likely that the macro- period of 1 5 8 A not only includes two myelin double layers but also structural protein (Schmitt, 1950b). Cf. Fig. 48, p. 57. TABLE XXXI THICKNESS OF BIMOLECULAR LAYERS OF LIPIDS IN NEURAL MYELIN (after BEAR, PALMER, AND SCHMITT, I941) Spacing in A Substance Determined by Calculated from atomic distances Lecithin . . . Cephalin . . . Sphyngomyelin Kerasin. . . . Phrenosin. . . 65 64 64 The myelin sheath does not entirely lose its birefringence when the myelin substances are extracted with fat solvents, but there then appears a negative cross on the cross-section (Schmidt, 1937a, b; Schmitt and Bear, 1939). This birefringence decreases appreciably when the extracted cross-sections of the nerves are transferred from alcohol to Canada balsam. There is therefore lamellar form bire- fringence, for the radial direction remains the optical axis, just as before extraction of the myelin. The submicroscopic layers must consist of neurokeratin, which is to be considered as the frame substance of the sheath. The polypeptide chains of this protein cannot have any preferred orientation, for, if they had, there would be no optical axis in the radial direction. The submicroscopic lamellae of protein must therefore be fohate in texture. Schmidt (1937a, p. 306, Fig. 80) assumes that there are individual submicroscopic particles of 362 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III protein which are to some extent independent of each other. This conflicts with the idea of these insoluble protein lamellae as frame substance. Fig. 179 represents the submicroscopic structure of the nerve sheath according to Schmidt (1937b). Neurokeratin lamellae running tangentially alternate with bimolecular lipid layers. It is difficult to say what the physiological significance of this foliate fine-structure may be. It should be noted that if this is destroyed, say by melting of the myelin substances, nerves lose their electric conductivity. a) ^3 b) Fig. 179. Fine-structure of medullated nerves, a) Optics. N neurofibrillar string, positively uniaxial as referred to the axial direction. M myelin sheath positively uniaxial as referred to radial direction (after Ambronn and Frey, 1926). h) Submicroscopic structure of the medullary sheath (after Schmidt, 1937b). A lamellae of protein. L bimolecular lipid layers. (Further details in F. O. Schmitt, 1936; O. Schmidt, 1942; v. MuRALT, 1946.) Another interesting fact is reported by Taylor (1942), who found that in nerves having approximately equal conduction velocities, the product of fibre diameter and sheath birefringence is roughly constant. The laminated fine-structure of the myelin sheath, found by in- direct methods, has been made visible in the electron microscope (Fernandez-Moran, 1950a, b). The periodicity of the lamination is 80 A, which is half the long-range X-ray diffraction period of 1 5 8 A reported above. Schmidt (1937a) detected a similar arrangement of lipid molecules orientated perpendicular to the parallel layers of protein in the outer members of the retinal cells in the eyes of Vertebrates, which has been PROTEINS 363 substantiated by electron micrographs (Sjostrand, 1949); and in this monograph (Fig. 151b, p. 259) such an arrangement has been shown to be probable in the microstructure of the chloroplasts. It looks, therefore, as if submicroscopic lamellar protein-lipid systems of the kind are fairly common in biological material. The axon. The protein of the nerve axon has been termed neuronin (Bear, Schmitt and Young, 1937). In the living nerve it constitutes only 3-4% of the fibre weight, the rest being an aqueous solution. r Fig. 180. Submicroscopic structure of an internodal segment of a myelinated nerve fibre (from Fernandez-MorAm, 1950, 1952a). N neurilemma, M myelin sheath, Ax axolemma, A axon, C collagen fibrils, E dark smooth fibrils. It is for this reason that X-ray absorption micrographs record 5 to 8 times less mass in the nerve axon (0.05 X io~'^'^ gj fji^) than in the myelin sheath (0.3 to 0.4 X lo'^^g^^s. Engstrom and Luthy, 1949, Engstrom and Lindstrom, 1950). This high dilution makes it un- certain whether the axoplasm exists as a sol or as a gel in the living state. Flaig (1947) reports that its viscosity is considerably increased during nerve activity, indicating a sol — gel equilibrium similar to that involved in protoplasmic flow. In fixed axoplasm, neurofibrils become visible. Their diameter ranges down to the resolving limit of the ordinary microscope. 364 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III However, in the electron microscope much finer filaments with 100-200 A diameter are visible (Fernandez-Moran, 1952b). Like other fibrous protein substances, the neurofibrils are posi- tively uniaxial, but their birefringence is very weak and is pushed into the background by the very strong anisotropy of the myelin sheath (Fig. 179a). The axoplasm shows form birefringence (Bear, Schmitt and Young, 1935). When heated, it shrinks lengthwise, like col- lagen fibres (Schmitt and Wade, 1935). The intrinsic birefringence of neuronin is 0.005, which is near to that of myosin (0.008). The mean refractive index as indicated by the minimum of the form birefringence curve amounts to 1.57-1.60^ a value which coincides with that of neurokeratin 1.58 and muscle myosin 1.576 (H, H. Weber, 1934). Fine-stnicture of nerves. Thin sections have yielded very instructive electron micrographs which settle several controversial points of nerve cytology. A distinct neurolemma which envelopes the myelin sheath is visible. At the nodes of Ranvier the axon is constricted but not intercepted (RozsA, Morgan, Szent-Gyorgyi and Wyckoff, 1950a, b). Fernandez-Moran (1950, 1952a) has compiled the results of his electron microscopic studies in a diagrammatic outline which is re- produced in Fig. 180. There is a 200 A thick granular neurolemma (N) with dark smooth fibres (E) which resemble elastic fibres, and adhering cross-striated collagen fibres (C). The sheath (M) consists of about 50 thin con- centric lamellae with an average periodicity of 80 A. The interlamellar spaces are locally inflated. The sheath is separated from the axon by a reticulate membrane, the axolemma (Ax), formed by beaded filaments 100-200 A in width. In the axis cylinder a verv fine reticulum is visible. g. Fibrillar Proteins. Recapitulation. The important frame proteins are of the fibrillar type. Their poly- peptide chains have a strong tendency to crystallize by forming chain lattices. X-ray diffraction studies have disclosed two types of axial spacings in these lattices, which have been classified as the keratin- myosin and the collagen group (Astbury, 1947; Marks, Bear and Blake, 1949). PROTEINS 365 3.5i Fig. 181. \^olume of an amino acid residue. Keratin-myosin group. In extended crystallizing polypeptide chains the space needed by an amino acid residue in the direction of the chain axis is 3.5 A. This spacing is called main chain spacing; the fibre period found by X-rays is usually a multiple of this value. The lateral distance of the main chains in the direction perpen- dicular to the plane of the side chains, termed backbone spacing by AsTBURY, is 4.5 to 4.6 A. In the third direction, the side chain spacing depends on the length of the radicals R of the amino acids involved. With the exception of silk fibroin, this spacing is astonishingly constant, indicating an average length of the amino acid residues of about 10 A. Therefore, the average volume of an amino acid residue is roughly 3.5 A X 4.6 A X 10 A = 161 (A)=^ (Fig. 181). In Mammals ectodermal formations such as hairs, feathers, epi- dermis, nails, horns and the mesodermal proteins muscle myosin, blood fibrinogen and fibrin correspond to this type. Astbury (1947) has therefore called it keratin-myosin-epidermis-fibrinogen (k-m-e-f) group. The special interest of this group is the possibility of the ^^^a transformation of its polypeptide chains (Fig. 171, p. 339), whereby the main chain spacing of three amino acid residues is reduced from 10.5 A to 5.1 A. Many natural fibrous proteins exist in this folded form (Table XXXII). As shown by Astbury, the a-chains can be reversibly transformed into the ^^-configuration ; hence they display an inherent elasticity and potential contractility. The muscle protein of Invertebrates is also of this type (clam muscle of Molluscs) and even the bacterial flagella of Proteus vulgaris show the characteristic spacing of 5.1 A (Astbury and Weibull, 1949). Silk fibroin differs from the fibrous proteins of the k-m-e-f group not only by its short side chain spacing but also by its lack of any /S ^ a transformation. The specificity of the different proteins in the k-m-e-f group is due to the special share of the different amino acids and their arrangement along the polypeptide chain. It is thought that bonds between the side chains stabilize the chain lattice both of the ^- and of the a-form. 366 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III > C < rt en ON ON cr^ ,^^ H4 ON t/1 ^ ,", u ..1 •^ ^ D O o Pi a: in e^ rr\ ^ ^ Q ■u C\ ^^ rrv r7\ 'U c t~< hH •+ c r rt ^^ ^-^ ^— ^ -^ c« >H i» (« >^ ^ G\ >< n1 k; a Bi ai Hi ^_^ ei N D D D D O D m m m CQ m a; n H F- H H H <- H c/l v: VI c/5 c/l HI c/l l^ <: < < <^ -< CQ < H W o\ O C\ ^ ^ ■=!- OS ON OS ^•^ »— t ■^ ,0^ ^Ph w w w °^ -li •^ !.d U. ^ k; <: < < „ D u J J z < 3^^ ffl CQ CQ -n X) 13 t! '^ r: C C -< c CJ ci « ai b; o; < < < W w w :« H r- m CQ CQ ASTBUR Champe f^ OS M V) f/5 N..^- a! a: Bi ^ ^ ^ § S « » 1^ ^ - c rt oj C C O -1 § ^ 2 « * * Z )-i w H o OS (1. )— H i~^ r/l X D X O X a CQ o H (/5 0 z hH u ^ 00 r-~ r~ i so NO i^^ I/-N \^-\ h-i *r\ \o O ^O VO w 00 00 oo r- o d On On On d i-< O d 1^ NO NO NO r- o ^NO ^ -1- Tt- 1- t -t NO •^s 00 f<^ CO MN oo On ►- I-- O «■ d " N ■-' 00 00 On OO c ^c c c . "U ^ tE ^ O '3 -^i •a E s .2 3 bfl 6 a n 9 « y CL, Ph U CQ 2 PROTEINS 567 When the side bridges of these chains are detached, the primary- valence chains crumple and contract with great force. Cystine-sulphur bridges, which only relatively drastic treatment can rupture, are sup- posed to be active in a-keratin. The side chain bonds in actomyosin, on the other hand, are far more labile, with the result that only a slight change in the reaction of the surrounding medium is needed for contraction. On account of its lability, actomyosin has been com- pared with raw rubber and keratin with vulcanized rubber and, as already stated, the tonofibrils have been described as "vulcanized" myofibrils (Mark and Philipp, 1937). Convenient as such com- parisons undoubtedly may be, they should be applied only with the utmost discretion, particularly as long as our knowledge of acto- myosin is no fuller than it is at the present time. Collagen group. The proteins of the collagen type are wide-spread in the animal kingdom. They occur as collagen proper in the meso- dermal tissues of Vertebrates (connective tissue, tendons, bones), as elastoidin in scales and fins of fishes, as ovokeratin in the egg capsule of rays, as ichthyocol in the swimbladder of fishes, as bysso-keratin in the byssus threads of Pinna nohilis (Mollusca), as ascaro-collagen in the cuticle of Ascaris (Nematoda), as connective tissue in the peristome of the sea urchin Arbacia (Echinodermata), in the axial stalk of the sea pens (Coelenterata), as spongin in the Porifera etc. (IVIarks, Bear and Blake, 1949). All these fibrous proteins have a main chain spacing of 2.8 A, whilst the backbone and side chain spacings are similar to those of the k-m-e-f group with the main chain spacing of 3.5 A (Table XXXII). The shortening of the collagen fibre spacing by 20% is due to the cis-position of the chain member >CHR related to the peptide bond -NH - CO- or -N = COH- (Fig. 173, p. 346). Thus the fibre proteins of the collagen group are composed of polypeptide chains in the cis-form, whilst those of the k-m-e-f group assume the trans-form. The latter are capable of re- versible contractions, whereas those of the collagen group have a strong tendency towards an irreversible supercontraction when the lateral bonds of the chain lattice are destroyed; as an extreme result, globular proteins can be formed (e.g. gelatin). Long-range spacings. Besides the short main chain spacings which characterize the keratin and the collagen group, there are long-range spacings in the fibre proteins which are disclosed by low angle X-ray 368 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III diffraction. The results of such studies are collected in Table XXXII. In silk fibroin Friedrich-Freksa, Kratky and Sekora (1944) found a period of 70 A, which corresponds to 20 amino acid residues. In feather keratin a somewhat longer spacing of 95 A is reported. In porcupine quill there is a long-range spacing of 198 A and, as its keratin is present in the a-form, where 3 amino acid residues cover 5.1 A, 116 residues (which is near to 2^ x 3^ = 108) would constitute such a period. In the adductor muscle of the mollusc Venus mercenaria a small-angle spacing of even 725 A has been reported, which seems to be divided into four subspacings of 145 A. In this spacing 426 amino acids can be lodged (which can be associated with 2^ X 3^ = 432). Bear (1944), who has measured these long fibre periods, discovered transverse long-range spacings as well; they amount to about 0.4 of the fibre spacings reported (Table XXXII). Consequently, there is not only a repetition of definite sequences of amino acids in the main chains, but at the same time distinct numbers of polypeptide chains are collected into crystallographic units. In contrast to this behaviour, which shows that the proteins of the keratin-myosin group exist in a three-dimensional crystalline state, BoLDUAN and Bear (1950) have only found a long-range fibre spacing in the collagen group but no transverse spacings. The inference is, therefore, that there is no true crystal pattern in the collagen proteins, but simply an arrangement of parallel chains, similar to that of liquid smectic crystals with only a unidirectional periodicity. The long-range fibre period seems to be the same for all collagen proteins investigated ; it measures 640 A (Marks, Bear and Blake, 1949) and agrees with the striation seen in the electron microscope, corresponding to 228 amino acid residues. There seems to be more uniformity in the proteins of the collagen group than in those of the keratin-myosin-epidermis-fibrinogen group. General occurrence of striated protein fibrils. Microscopic histology considered the striated muscle fibrils as a special case of protein fibres. The electron microscope has, however, revealed the fact that banding is a general feature of fibrillar proteins, the period of this striation being submicroscopic. It has been found in smooth muscle fibrils (Hall, Jakus and Schmitt, 1945), collagen fibrils (Fig. 174, p. 349), precipitated blood fibrin (Wolpers and Ruska, 1939), ejected tricho- ■cysts oi Paramecium (Jakus, 1945; Wohlfahrt-Bottermann, 1950; Z PROTEINS 569 Knoch and Konig, 195 i) etc. In collagen fibres the striation period of 640 A corresponds to the long-range spacing discovered by small- angle X-ray diffraction. This method discloses even in keratin a long- range periodicity of 200 A (McArthur, 1945). At first sight this widespread occurrence of a submicroscopic striation in fibrous proteins seemed rather enigmatic. But its formation can be studied nowadays, since there are soluble proteins which yield striated fibrous proteins on precipitation. Such an example is blood fibrin. Even more interesting is the fact that dissolved collagen can be reconstituted to precipitated collagen fibrils with a striation period of 635 A (Bahr, 1950). These experiments favour the view that the striated microfibrils are formed by linear aggregation (Fig. 104a, p. 160) of globular particles. In this way the submicroscopic striation is easily understood, but it is difficult to explain how a chain lattice with polypeptide chains very much longer than the diameter of the dissolved protein particles is formed. In this dilemma a helpful suggestion may be that in globular proteins the amino acids are only loosely bound and not yet tied together by firm peptide bonds (see p. 329). Then, on denaturation by precipitation, not only should peptide bonds be formed inside the globular protein macromolecule, but should also bridge the amino acids of the adjacent molecule, the result being polypeptide chains running straight through numerous protein particles. It is more likely that some such mechanism is involved than that preformed poly- peptide chains curled up in the globular particle should unfold com- pletely to form straight threads, which would be necessarily entangled before a chain lattice can be formed. Chemical changes of the protein molecule due to the transformation globular -> fibrillar of its shape have been recorded in fibrinogen (Bailey, Bettelheim, Lorand and Middlebrook, 195 i). When blood clots, fibrinogen (M.W. 500,000) is transformed into fibrillar fibrin by the enzyme thrombin. This change is associated with the appearance of amino-terminal residues of glycine by specific hy- drolysis. Whereas fibrinogen has no such end groups, five terminal glycine residues appear per mole of fibrinogen when converted into fibrin. It should be emphasized that ordinary denaturation does not cause this effect and that only thrombin is capable of inducing it. 370 FINE-STRUCTURE OF PROTOPLASMIC DERIVATIVES III Conclusion Whatever the final explanation of these important molecular trans- formations may be, the typical properties of the polypeptide chains may be said to be the general tendency to agglomerate into fibrous strands and their widespread poiver of contracting (actomyosin, keratin, collagen). Thus the very structure of protoplasmic polypeptides furnishes the fundamental conditions for fibrillar differentiation and contractility. RETROSPECT A revolutionary fact which emerged from the synthesis of organic compounds was that, in chemistry, there is no fundamental difference between living and inanimate matter. The complicated process of metabolism is not controlled by some special vital principle, but has its being in the co-ordination of innumerable reactions, each and all, being separately accessible to causal investigation. Yet no simple mechanistic interpretation can account for their delicately attuned harmony and their purposiveness. Morphological formations in the submicroscopic world present an exactly similar case. Whoever had expected to find special formative principles, alien to the inanimate world, in these invisible regions, is doomed by the results of research into natural substances of high molecular weight to as great a dis- appointment as was at one time suffered by the believers in mysterious life forces which alone were deemed capable of building up organic compounds. The formative forces in protoplasm and its derivatives are no different from those operating within inanimate Nature. There is no evidence of the existence of formative principles beyond the atomic valency and the various molecular cohesive forces in their vari- ous patterns. This need cause no surprise if it be remembered that, in the molecular world, the chemical and formative properties merge into each other. In that realm, chemistry and morphology become in- separably one, since every morphological change which a molecule undergoes inevitably involves chemical changes. All metabolic processes therefore run parallel to changes in molecular form. For this reason substance and form are closely interrelated, not only in the inanimate world, where every compound can be clearly classified by its molecular or crystal structure, but in living matter as well. The idea of an essential difference between the morphology of the animate and that of the inanimate world has no place in the theory of sub- microscopic morphology. Just as organic chemistry grew out of inorganic chemistry and has 37^ RETROSPECT its roots in the fundamental principles of the latter, so should bio- morphology be considered simply as a highly developed system, evolved from molecular and micellar morphology to the shaping of cells and organisms. Only the first step in this development at present lends itself to deductive reasoning, viz., the transition from molecular to micellar morphology. This has been made possible by the modern evidence on the structure of highly polymeric chain molecules and globular macromolecules. There are two guiding principles, of the utmost importance to biomorphology, which are already recognizable in the configuration of chain molecules. They are: i. The principle of repetition, which is the foundation of all lattice structures and of every form of banding, and 2. The principle of specificity. The first principle is represented, on the one hand, by the ever-recurring members of the chain (intra- molecular spacing) and, on the other, by the assemblage into a lattice pattern of kindred chains (intermolecular spacing), as for example frame substances, reserve substances, and lipid layers. Only if all the members of a certain kind of chain are of exactly the same structure can true intermolecular repetition take place. This law does not normally apply to polypeptide chains, since their side groups are often of different structure. In consequence, we find the second principle holding sway, i.e., the capacity of otherwise similar molecular ele- mentary units to assume a specific arrangement which may be repeated for its part in long-range periods. We do not yet know how the visible specific forms of cellular organelles, cells, tissues and organ- isms grow out of this specificity, but doubtless causal relations do ■exist between molecular morphology and morphogenesis, as fore- shadowed by enzyme chemistry and the asymmetrical synthesis of organic compounds. A problem no less difficult than causal morphological development is that of the molecular morphology of heredity; for, assuming that €very kind of visible form owes its origin to particular configurations of concrete hereditary entities which cannot arise spontaneously, then their complicated structure must be constantly reproducing their like. Although the multiplication of the virus molecules presents some analogy to this, we have nothing to go upon to build up a clear picture of the auto-reproduction of those complicated structures, the genes. Fox the present, submicroscopic morphology has been successful RETROSPECT 573 Only in so far as specificity is ignored, but within this modest sphere the knowledge acquired is most significant. The substratum in which life is inherent is not a disperse phase with individual particles or ultramicrons ; it possesses 2l structure. Its active centres, which control development, are arranged in a given order. They are not intermingled by mere laws of chance and Brownian molecular movement ; the fact is rather that they arrange themselves into a delicate, very plastic and flexible pattern, actuated, as it were, by a purposeful, co-ordinative impulse. No more than leaves, blown by autumnal winds from the twig and fluttering helplessly in the air, are able to assimilate for the parent tree, can independent, ambulant, reactive molecules take part in any organized work. It is not surprising, therefore, that the active groups of the enzymes should only be capable of acting in association with a carrier of a given structure. For, orderly biological processes are unthinkable without presupposing structure, and it is therefore out of the question that any living constituent of protoplasm could consist of structureless, fluid, independently displaceable particles. It is for this reason that colloid chemistry, based, as it is, upon the disperse principle, has thrown so little light upon the submicroscopic structure of protoplasm. For the cell certainly is not a pouch filled with ultramicrons suspended in a fluid, whirling about haphazardly and in confusion; it is, on the contrary, a wonderful system, the intrinsic structure of which, could it but be seen, would assuredly fill every observer with an enthusiasm equal to that which microscopic cytomorphology inspires. It is true that metabolic centres (lyoenzymes, mitochondria, ery- throcytes, chloroplasts) are independent of each other; but their movement does not obey the law of entropy; they are actively directed to the localities where their biochemical capacity is needed. On the other hand, the special cytological and histological systems which facilitate an appropriate production and distribution of those metabolic centres (protoplasmic flow, blood capillaries, glands) must have some coherent structures at their disposal. The organization of these semi-solid structures is responsible for the creation of biological objects of any shape or form and, therefore, is the very foundation of morphogemsis. In the inanimate world, crystallization will at times produce structures from an amorphous mass; but the structures of living 374 RETROSPECT protoplasm cannot be spontaneous!}" generated from unformed solutions because, complicated and delicately inter-adjusted as they are, they can only actualize in contact with already existing structures. Hence the supreme axiom of cytology, namely, that all cells derive from their like, applies equally, though in a wider sense, to invisible, submicroscopic cytogenesis : STRUCTURA OMNIS E STRUCTURA I LITERATURE Abbe, E., 1879. (Microscopy, immersion objective.) S. B. Jenaische Z. Medizin u. Natur- wiss. 10. Jan. 1879. .\dam, N. K., 1930. The physics and chemistr}- of surfaces. Oxford. 3rd edit. 1941. Algera, L., Beijer, J. J., v.\N Iterson, W., Karstens, W. K. H. and Thung, T. H., 1947. (Chloroplast, electron microscopy.) Biochim. Biophys. Acta i. 517. Ambronn, H., 1888. (Cuticular layers, optics.) Ber. dtsch. hot. Ges. 6. 226. Ambronn, H., 1889. (Cherr)' gum, optics.) Ber. dtsch. hot. Ges. 7. 103. Ambronn, H., 1891. (Freezing of colloids.) Ber. Sachs. Akad. Wiss., math.-phys. Kl. 43. 28. Ambronn, H., 1916/17. (Rodlet double refraction.) Kolloid-Z. 18. 90, 273. 20. 173. Ambronn, H., and Frey, A., 1926. Das Polarisationsmikroskop. Leipzig. Anderson, D. B., 1928. (Cuticular layers, microchemistr\'.) Jahrb. wiss. Bot. 69. 501. Anderson, D. B. and Kerr, Th., 1938. (Cotton, crossed systems.) Industr. Engng Chem. 30. 48. Anderson, D. B. and Moore, J. H., 1937. (Cotton hairs without growth layers.) Amer. J. Bot. 24. 503. Andresen, N., 1942. (Cytoplasmic compounds.) C. R. Lab. Carlsberg ser. chim. 24. 140. Ardenne, M. von, 1940a. Elektronen-Ubermikroskopie. Berlin. Ardenne, M. von, 1940b. (Results of electron microscopy.) Naturvviss. 28. 113. Arisz, W. H., 1943. (Active and passive resorption.) Versl. Ned. Akad. Wetensch. Afd. Natuurk. 52. 639. .\risz, W. H. and \'an Dijk, P. J. S., 1939. (Asparagine intake by Vallisneria.) Proc. Acad. Sci. Amsterdam 42. 820. Arnon, D. L and Stout, P. R., 1939. (Molybdenum, indispensable element for plants.) Plant Physiol. 14. 599. Aronoff, S., 1950. (Chlorophyll, review.) Bot. Rev. 16. 525. AscENZi, A., 1950. (Bones, form birefringence.) Science 112. 84. Astbury, W. T., 1933a. (Polypeptide chains, periodicity.) Trans. Faraday Soc. 29. 195. AsTBURY, W. T., 1933b. (Silk fibroin, side chains.) J. Soc. Dyers Col. Bradford 49. 168. Astbury, W. T., 1933c. Fundamentals of fibre structure. Oxford. Astbury, \X'. T., 1936. (Lactam-lactim transformation.) J. Text. Inst. Manchr. 27. 282. Astbury, W. T., 1940. (Collagen chain molecule.) J. int. Soc. Leath. Chem. 24. 69. Astbury, W. T., 1947. (k-m-e-f proteins.) Proc. Roy. Soc. London B 134. 303. Astbury, W. T., 1947/49. (Muscle, X-ray analysis). Exp. Cell Research Suppl. I, Stock- holm, p. 234. Astbury, W. T. and Atkin, W. R., 1933. (Gelatin, X-ray analysis). Nature 132. 348. Astbury, W. T. and Bell, F. O., 1938. (Nucleic acids, chain lattice.) Nature 141. 747. Astbury, W. T. and Bell, F. O., 1941. (a-Keratin, intramolecular fold.) Nature 147. 696. Astbury, W. T. and Dickinson, S., 1935a. (Myosin, a-/3-transformation.) Nature 135. 95. Astbury, W. T. and Dickinson, S., 1935b. (Muscle fibre, a-/3-transformation.) Nature 135. 765. Astbury, \V. T., Dickinson, S. and Bailey, K., 1935. (Edestin and ovalbumin, X-ray analysis.) Biochem. J. 29. 2351. 576 LITERx\TURE AsTBURY, W. T, and Lomax, R., 1934. (Pepsin, amorphous.) Nature 133. 795. AsTBURY,W. T. andMARWiCK.T. C, 1932. (Keratin of feathers.) Trans. Faraday Soc. 29. 206. AsTBURY, W. T. and Weibull, C, 1949. (Protein of bacterial flagella.) Nature 163. 280. AsTBURY, W. T. and Wrinch, D. H., 1937. (Polypeptide chains, keto- enol- tautomery.> Nature 139. 798. Baas Becking, L. G. M. and Koning, H. C, 1934. (Chlorophyll, spectrum.) Proc. Acad. Sci. Amsterdam 37. 674. Badenhuizen, N. p., 1937. (Starch grains, microscopic structure.) Protoplasma 28. 293. Badenhuizen, N. p., 1938. (Starch grains, microscopic structure.) Protoplasma 29. 246. Bahr, G., 1950. (Reconstitution of collagen fibrils.) Exp. Cell Research i. 603. Bailey, K., 1942. (Muscle protein.) Advances in Protein Chem. i. 289. Bailey, K., Bettelheim, F. R., Lorand, L. and Middlebrqok, W. R., 195 i. (Fibrinogen,. clotting.) Nature 167. 233. Baltzer, F., 1942. (Morphogeny.) Verh. schweiz. naturf. Ges. Sion 1942, p. 59. Bank, O., 1941. (Nucleus as complex coacervate.) Protoplasma 35. 419. Barlund, H., 1929. (Permeability.) Acta bot. fenn. 5. i. Bartholome, E., 1936. (Cohesion forces in liquids.) 24. 824. Bates, F. L., French, D. and Rundle, R. E., 1943. (Starch, amylose content.) J. Amer. chem. Soc. 65. 142. Baud, Ch. A., 1949a. (Nuclear membrane, chemical composition.) Bull. Hist. appl. 1949^ No. I, p. 12. Baud, Ch. A., 1949b. (Nuclear membrane, double refraction.) Bull. Hist. appl. 1949,. No. 5, p. 99. Beams, H. W., 1943. (Cytoplasmic components.) Biol. Symp. 10. 71. Beams, H. W., Evans, T. C, Van Breemen, V. and Baker, W. W., 1950, (Mitosis, electroa microscopy.) Proc. Soc. exp. Biol. 74. 717. Bear, R. S., 1944. (Collagen, keratin. X-ray diffraction.) J. Amer. chem. Soc.66. 1297,2045. Bear, R. S. and French, D., 1941. (Starch grains. X-ray diffraction.) J. Amer. chem. Soc. 63. 2298. Bear, R. S., Palmer, K. J. and Schmitt, F. O., 1941. (Nerve lipids. X-ray analysis.) J. c;ll. comp. Physiol. 17. 357. Bear, R. S. and Schmitt, F. O., 1936. (Nerve myelin, optics.) J. opt. Soc. Amer. 26. 206. Bear, R. S., Schmitt, F. O. and Young, J. Z., 1937. (Nerve axoplasm.) Proc. Roy. Soc. London 123. 505. Becker, W. A., 1934. (Youngest cell wall.) Acta Soc. Bot. Polon. 11. 139, 190. Becker, W. A., 1935. (Youngest cell wall.) Z. Zellforsch. 23. 253. Becker, W. A., 1936. (Vital staining.) Protoplasma 26. 439. Becker, W. A. and Kozbial,J., 1937. (Chromosomes, double refraction.) Acta Soc. Bot. Polon. 14. 239. Belar, a., 1929. (Nucleus, spindle.) Arch. Entw. Mech. Org. 118. 359. Bensley, R. R., 1937. (Mitochondria as coacervates.) Anat. Rec. 69. 341. Bensley, R. R., 1938. (Plasmosin.) Anat. Rec. 72. 351. Bensley, R. R., 1943. (Chemistry of cytoplasm.) Biol. Symp. 10. 323. Bensley, R. R. and Hoerr, N. L., 1934. (Mitochondria, chemistry.) Anat. Rec. 60. 449. Berger, C. a., 1941. (Endomitosis.) Cold Spr. Harb. Symp. quant. Biol. 9. 19. Bergmann, M. and Niemann, C, 1937. (Silk fibroin.) J. biol. Chem. 118, 301; 122. 577. Bernal, J. D., 1932. (Sterins, X-ray diffraction.) Chem. Industr. Berl. 51. 466. Bernal, J. D., 1939. (Virus, intramolecular interferences.) Proc. Roy. Soc. London A 170. 74. Bernal, J. D., 1940. The cell and protoplasm, p. 199. Washington D. C. Bernal, J. D. and Crowfoot, D., 1934. (Pepsin, X-ray diffraction.) Nature 133, 794. LITERATURE 377 Bernal, J. D. and Fankuchen, I., 1937. (Cr}'stalli2ed virus, X-ray diffraction.) Nature 139. 925. Bernhard, W., Gautier, A. and Oberling, Ch., 195 1. (Liver cells, beaded chains.) C. r. Soc. Biol. Paris 145. 566. Bersin, Th., 1939. Kurzes Lehrbuch der Enzymologie. Leipzig. Berthold, G., 1886. Studien uber Protoplasmamechanik. Leipzig. Bessis, M., and Bricka, M., 1948. (Cytoplasm, electron microscopy.) Biochim. Biophys. Acta 2. 339. Bessis, M. and Bricka, M., 1949. (Erythrocyte, electron microscopy.) Arch. Anat. micr. 38. 190. Biedermann, W., 1903. (Solid secretions.) Z. allg. Physiol. 2. 395. BiNZ, A., 1892. (Chloroplast, grana structure.) Diss. Univ. Zurich; Flora 76. Erg.bd. zu Jg. 1892. Blackmann, F. F., 1905. (Blackmann's dark reaction.) Ann. Bot. Lond. 19. 281. Bladergroen, W., 1945. Physikalische Chemie in Medizin und Biologic. Basel. Blank, F. and Frey-Wyssling, A., 1941. (Cell elongation.) Ber. schweiz. bot. Ges. 51. 116. Bl.\nk, F. and Frey-Wyssling, A., 1944. (Protoplasmic growth.) Ann. Bot. Lond. 8. 71. Block, R. J., 1939. (Wool, ratio of amino acids.) J. biol. Chem. 128. 181. BoEHM, G., 1933. (Ner^^e fibres. X-ray diffraction.) Kolloid-Z. 62. 22. Boehm, G., 1939. (Double refraction of flow, method of investigation.) Handb. biol. Arbeitsmeth. v. Abderhalden, Abt. II, Teil 3, p. 3939. Boehm, G. and Weber, H. H., 1932. (Myosin, X-ray diffraction.) Kolloid-Z. 61. 269. Bogen, H. J., 1938. (Permeability to urea and glycerol.) Planta 28. 535. BoLDUAN, O. E.A. and Bear, R. S., 1950. (Collagen, long-range spacing.) J. Polymer Sci. 5. 159. Bonner, J., 1935. (Cell elongation.) Jahrb. wiss. Bot. 82. 377. Bonner, J., 1936a. (Pectins.) Bot. Rev. 2. 475. Bonner, J., 1936b. (Respiration of Avena coleoptile.) J. gen. Physiol. 20. i. Borries, B. von, 1941. (Electron microscopy.) Stahl u. Eisen 61. 725. Borries, B. von and Ruska, E., 1939a. (Electron microscopic image.) Naturwiss. 27. 281. Borries, B. von and Ruska, E., 1939b. (Electron microscope.) Z. w'ss. Mikr. 56. 517. BossHARD, H. H., 1952. (Cell walls, growth in area.) Diss. E. T. H. Zurich. BoT, G. M., 1939. (Chloroplasts, proportion of P.) Diss. Leiden. BoTTiNi, O., 1937. (Intramicellar base exchange.) Kolloid-Z. 80. 56. Bougault and Bourdier, 1908. (Etholids.) C. R. Acad. Sci. Paris 147. 1311. Bourne, G., 1945. (Mitochondria and Golgi apparatus.) Cytology and cell physiology, p. 99. Oxford. Brandenberger, E. and Schinz, H. R., 1945. (Bones, hydroxyl-apatite.) Helv. med. Acta A 12. Suppl. XVI. Brauner, L., 1930. (Cuticular layers, electrical charge.) Jahrb. wiss. Bot. 73. 513. Brauner, L., 1932. (Swelling of gels by ions.) Pflanzenphysiol. Praktikum II. Teil, p. no. Jena. Brauner, L. and M., 1943. (Water resorption depending on respiration.) Rev. Fac. Sci. Univ. Istanbul B 8. 30. Brenner, S., 1949. (Mitochondria, redox system.) S. Afr. med. Sci. 14. 13. Bretschneider, L. H., 1948. (Sperm, fine-structure.) Vakbl. voor Biologen 28. 137. Bretschneider, L. H., 1949a. (Electron microscopy, thin sections. Rocking microtome.) Proc. Conf. Electron Micr. Delft. Bretschneider, L. H., 1949b. (Bull sperm, electron microscopy.) Proc. Acad. Sci. Amsterdam 52. 301. Bretschneider, L. H., 1950a. (Ciliates, electron microscopy.) Mikroskopie (Wien) 5.257. 378 LITERATURE Bretschneider, L. H., 1950b. (Sperms, electron microscopy.) Proc. Acad. Sci. Amster- dam 53. 531. Bretschneider, L. H., 1950c. (Differently fixed cytoplasm, electron microscopy.) Proc. Acad. Sci. Amsterdam 53. 1476. Brill, 1941. (Natural silk, action of spinning.) Naturwiss. 29. 221. Brill, R., 1943. (Silk fibroin.) Z. phys. Chem. B 53. 61. Brown, D. E. S., 1934a. (Cytoplasm, high pressure experiments.) J. cell. comp. Physiol. 4- 257- Brown, D. E. S., 1934b. (Viscosity of plasm depending on pressure.) J. cell. comp. Physiol. 5- 335- Brown, G. L., Callan, H. G. and Leaf, G., 1950. (Nuclear sap, chromatography.) Nature 165. 600. Brown, H. P., 1945. (Hairy flagella, electron microscopy.) Ohio J. Sci. 45. 247. Bull, H. B., 1947. (Protein, monolayers.) Advances in Protein Chem. 3. 95. BuNGENBERG DE JoNG, H. G., 1932. (Coaccrvation.) Protoplasma 15. no. Bungenberg, de Jong, H. G. and Bonner, J., 1935. (Films of phosphatides.) Protoplas- ma 24. 198. Burstrom, H., 1942. (Cell elongation, osmotic conditions.) Lantbrukshogskolans Anna- ler 10. I . Burstrom, H., 1951. (Cell elongation.) Physiol. Plantarum 4. 199. Burton, E. F. and Kohl, W. H., 1946. The electron microscope. New York. BiJssEM, W. and Weyl, W., 1936. (Glass, amicroscopic structure.) Naturwiss. 24. 324. Callan, H. G., Randall, J. T. and Tomlin, S. G., 1949. (Nuclear membrane.) Nature 163. 280. Callan, H. G. and Tomlin, S. G., 1950. (Nuclear membrane.) Proc. Roy. Soc. London B 137. 367. Caspersson, T., 1936. (Chromosomes, ultra violet absorption.) Skand. Arch. Physiol. Suppl. 8 of vol. 73. Caspersson, T., 1940a. (Nucleus, distribution of proteins). Chromosoma i. 562. Caspersson, T., 1940b. (Chromomeres, distribution of nucleic acids.) Chromosoma i. 605. Caspersson, T., 1941. (Protein metabolism in the cell.) Naturwiss. 29. 33. Caspersson, T. and Thorell,B., 1941. (Muscle, adenyl nucleotide.) Naturwiss. 29. 363. Castle, E. S., 1936. (Chitin, rodlet double refraction.) J. gen. Physiol. 19. 797. Castle, E. S., 1937a. (Spiral growth in Phycomyces.) J. cell. comp. Physiol. 9. 477. Castle, E. S., 1937b. (Cell wall extension.) J. cell. comp. Physiol. 10. 113. Castle, E. S., 1942. (Spiral growth in Phycomyces.) Amer. J. Bot. 29. 664. Chambers, R., 1925. (Microdissection.) Amer. J. Anat. 35. 385. Chambers, R., 1928. (Microinjection.) Biol. Bull. 55. 369. Champetier, G. and Faure-Fremiet, E., 1937. (Elastoidin fibres.) J. chim. phys. 34. 197. Champetier, G., and Faur£-Fremiet, E., 1938. (Collagen, X-ray analysis.) J. chim. phys. 35. 223. Chargaff, E., 1950. (Nucleic acids, specificity.) Experientia 6. 201. Chibnall, a. C, 1939. Protein Metabolism in the Plant. New Haven and London. Chibnall, a. C, Piper et al., 1934. (Plant waxes.) Biochem. J. 28. 2175, 2189. Chodat, R., 1907. Principes de Botanique. Geneve. Cholodny, N. and Sankewitsch, E., 1933. (Action of ions on protoplasm.) Protoplasma 20, 57. Christiansen, G. S. and Thimann, K. V., 1950. (Cell wall, protein content.) Arch. Biochem. 26. 230. Clark, G. L. and Corrigan, K. E., 1932. (Insulin, X-ray diffraction.) Phys. Rev. 40. 639. Claude, A., 1944. (Mitochondria, constitution.) J. exp. Med. 80. 19. LITERATURE 379 Claude, A., 1946. (Mitochondria, isolation by centrifugation.) J. exp. Med. 84. 51. Claude, A., and FuLLAM.E.F., 1 945. (Mitochondria, electron micrograph.) J. exp. Med. 81. 5 1. Claude, A. and Fullam, E. F., 1946. (Liver sections, rotating high speed microtome.) J. exp. Med. 83. 499. CoHN, E. J. and Edsall, J. T., 1943. Proteins, amino acids and peptides. New York. Collander, R., 1932. (Lipid filter theon,'.) Handb. d. N'aturwiss. Jena. 2 Aufl. 7. 804. Collander, R., 1937a. (Lipid filter theory-.) Schr. phys.-okon. Ges. Konigsb. 69. 251. Collander, R., 1937b. (Permeabilit^^) Ann. Rev. Biochem. 6. i. Collander, R. and holmstrom. A., 1937. (Adenoid action of cytoplasm.) Acta Soc. Fauna Flora fenn. 60. 129. Comar, C. L., 1942. (Chloroplast, chemical composition.) Bot. Gaz. 104. 122. Correns, C, 1893. (Cell walls of algae.) Zimmermanns Beitr. Morphol. Physiol. Pflanzen- zelle I. 260. Crick, F. H. C. and Hughes, A. F. W., 1950. (Cytoplasm, magnetic particle method.) Exp. Cell Research i. 37. Crowfoot, D., 1958. (Cn-stallized insulin, Patterson - Fourier analysis.) Proc. Roy. Soc. London A 164. 580. Crowfoot, D., 1939. (Protein crystals, X-ray analysis.) Proc. Roy. Soc. London A 170. 74. Crowfoot, D., 1941. (Cr\'stals of insulin, horse methaemoglobin and lactoglobulin.) Sci. Chem. Rev. 28. 215. Curtis, H. J., 1936. (Plasmalemma.) J. gen. Physiol. 19. 929. CzAjA, A. Th., 1930. (Ultrafiltration.) Planta 11. 582. Danielli, J. F., 1936. (Plasmalemma.) J. cell. comp. Physiol. 7. 393. Danielli, J. F., and Harvey, E. N., 1935. (Plasmalemma.) J. cell. comp. Physiol. 5. 483. Danon, D. and Kellenberger, E., 1950. (Microtome for electron microscopy.) Arch. Sci. Geneve. 3. 169. Davson, H. and D.anielli, J. F., 1943. The permeability of natural membranes. Cam- bridge. Denny, F. E., 1935. (Ethylene, stimulating substance.) Contr. Boyce Thompson Inst. 7- 97- Derksen, J. C, Heringa, G. C, and Weidinger, A., 1937. (Keratin, cornification.) Acta Neerl. Morphol. i. 31. Dervichian, D. G., Fournet, G. and Guinier, A., 1947. (Erythrocyte.) C. r. Acad. Sci. Paris 224. 1848. Dettmer, N., Neckel, J. and Ruska, H., 195 i. (Collagen fibres, electron microscopy.) Z. wiss. Mikr. 60. 291. Deuel, H., 1943. (Pectins.) Diss. E. T. H. Zurich. Deuel, H., 1947a. (Pectin, artificial cross-links.) HabiUtiationsschrift E. T. H. Zurich. Deuel, H., 1947b. (Esterification of polygalacturonic acid.) Experientia. 3. 151. Devaux, H., 1935a. (Films of albumin.) C.r. Soc. Biol. Paris 119. 1124. Devaux, H., 1935b. (Films of albumin.) C.r. Acad. Sci. Paris 200. 1560, 201. 109. DiEHL, J.M. andVAN-lTERSON, G., 1935. (Chitin, rodlet double refraction.) Kolloid-Z. 73. 142. DoBRY, A., 1938. (Coacervation.) J. Chim. phys. 35. 387. DoBRY, A., 1940. (Coacervation.) Bull. Soc. Chim. biol. 22. 75. Doutreligne, J., 1955. (Chloroplasts, grana.) Proc. Acad. Sci., Amsterdam 38. 886. Drawert, H., 1937. (Fixed tissues, staining.) Flora 32. 91. Drawert, H., 1948. (Permeability to urea.) Planta 35. 579. Drucker, B. and Smith, S. G., 1950. (Silk, paper chromatography.) Nature 165. 196. Ebner, V. VON, 1894/96. (Collagen, optics.) S.B. Akad. Wiss. Wien 103. 162, 105. 17. Elod, E.,NowoTNY,H.andZAHN,H., 1940a. (Wool, chemical reactions.) Kolloid-Z. 93. 50. 380 LITERATURE Elod, E., NowoTNY, H. and Zahn, H., 1940b. (Wool, fine-structure.) Melliand Textilber. No. 8. Elvers, I., 1943. (Chromosomes, electron micrograph.) Acta Horti Berg. 13. 149. Emerson, R. and Arnold, W., 1932. (Unit of assimilation ) J. gen. Physiol. 16. 191. Emerson, R. and Lewis, C. M., 1939. (Photosynthesis.) Amer. J. Bot. 26. 808. Engstrom, A. and Lindstrom, B., 1950. (X-ray absorption of microscopic objects.) Biochim. Biophys. Acta 4. 351. Engstrom, A. and Luthy, H., 1949. (Nerve, X-ray absorption.) Experientia 5. 244. Engstrom, A. and Ruch, F., 195 i. (Giant chromosomes, mass distribution.) Proc. nat. Acad. Sci. Wash. 37. 459. Euler, H. von, Bergman, B. and Hellstrom, H., 1934. (Chloroplasts of Elodea, chloro- phyll concentration.) Ber. dtsch. bot. Ges. 52. 458. Eymers, J. G. and Wassink E.G., 1938. (Purple sulphur bacteria, COj assimilation.) Enzy- mologia 2. 258. Fajans, K., 1923. (Deformation of ions.) Naturwiss. 11. 165. Fajans, K., 1925. (Deformation of ions.) Z. Kristallogr. 61. 18. Fankuchen, J., 1934. (Pepsin, X-ray diffraction.) J. Amer. Chem. Coc. 56. 2398. Fankuchen, J., 1941. (Proteins, X-ray analysis.) Cold Spr. Harb. Symp. quant. Biol. 9. 198. Farr, W. K. and Eckerson, S.H., 1934. (Cotton hairs, dermatosomes.) Contr. Boyce Thompson Inst. 6. 189, 309. Faure-Fremiet, E., 1936. (Elastoidin fibres.) Arch. Anat. micr. 32. 249. Faure-Fremiet, E., 1946. (Cytology, review 1940-1946.) Anne biol. 22. 57. Faure-Fremiet, E., Bessis, M. and Thaureaux, J., 1948. (Hyaloplasm, electron micros- copy.) Microscopic (Paris) i. 41. Felix, K., 195 i. (Frankfurt a. M.) Oral communication. Fernandez-Moran, H., 1950a. (Myelinated nerve, fine- structure of sheath.) Exp. Cell Research i. 143. Fernandez-Moran, H., 1950b. (Myelinated nerve, electron microscopy.) Experientia 6. 339. Fernandez-Moran, H., 1952a. (Myelinated nerve fibre.) Inaugural Diss. Uppsala. Fernandez-Moran, H., 1952b. (Axon of nerve fibre.) Exp. Cell Research 3. i. Feulgen, R. and Rossenbeck, H., 1924. (Nucleal staining.) Hoppe Seyler Z. physiol. Chem. 135. 203. Fischer, A., 1894. (Hairy flagella.) Jahrb. wiss. Bot. 26. 187. Fischer, F. G., 1942. (Nucleic acids, molecular structure.) Naturwiss. 30. 377. Fischer, H. et al., 1935. (Chlorophyll, molecular structure.) Liebigs Ann. 519. 209, 520. 88. Fitting, H., 1927. (Protoplasmic flow by hi .tidine.) Jahrb. wiss. Bot. 67. 427. Fitting, H., 1936. (Protoplasmic flow by histidine.) Jahrb. wiss. Bot. 82. 613. Flaig, J. v., 1947. (Nerve axon, colloidal state of neurin.) Neurophysiol. 10. 211. Foster, E., Baylor, M. B,. Meinkoth, N. A. and Clark, G. L. 1947. (Hairy flagella. electron microscopy.) Biol. Bull. 93. 114. Foster, F. J. and Samsa, E. G., 1950. (Ovalbumin, birefringence of flow.) Science 112. 475. Franck, J., 1935. (Theory of COg assimilation.) Naturu'iss. 23. 226. Frank, R., 1950. (Teeth, fine-structure of enamel.) Rev. mens. Suisse Odontol. 60. H09. Freudenberg, K., 1933. Stereochemie, Wien. Freudenberg, K., Schaaf, E., Dumpert, G. and Ploetz T., 1939. (Starch molecule, spiral structure.) Naturwiss. 27. 850. Freundlich, H., 1922. Kapillarchemie, Leipzig. Freundlich, H., 1937. (Properties of gels.) J. phys. Chem. 41. 901. Freundlich, H., 1942. (Thixotropy.) In W. Seifriz, The structure of protoplasm. Ames-Iowa, p. 85. LITERATURE . 381 Freundlich, H. and Lindau, C, 1932. (Mechanochemistty.) Handb. d. Naturwiss. v. Abderhalden, 2. Aufl. 6. 831. Freundlich, H., Stapelfeldt, F. and Zocher, H., 1924. (Double refraction of flow.) Z. phys. Chem. 114. 161, 190. Frey, a., 1924. (Types of double refraction.) Kolloidchem. Beih. 20. 209. Fret, A., 1925. (Dichroism of fibres, microtechnics.) Z. wiss. iVIikr. 42. 421. Frey, A., 1926a. (Micellar theory.) Ber. dtsch. bot. Ges. 44. 564. Frey, A., 1926b. (Cell walls, submicroscopic structure.) Jahrb. wiss. Bot. 65. 195. Frey, A., 1926c. (Closterium, viscosity of cell sap.) Rev. gen. Bot. 38. 273. Frey, a., 1927a. (Aspergillus, sporangiophore, optics.) Rev. gen. Bot. 39. 277. Frey, A., 1927b. (Dichroism of fibres.) Jahrb. wiss. Bot. 67. 597. Frey, A., 1928a. (School of Ambronn.) Kolloid-Z. 44. 6. Frey, A., 1928b. (Micellar science.) Protoplasma 4. 139. Frey, R., 1950. (Chitin of fungi.) Diss. E. T. H. Zurich 1950; Ber. schweiz. bot. Ges. 60. 199. Frey-Wyssling, a., 1930. (Micellar textures.) Z. wiss. Mikr. 47. i. Frey-Wyssling, a., 1932. (Latex tubes.) Jahrb. wiss. Bot. 77. 560. Frey-Wyssling, A., 1935a. Die Stoffausscheidung der hoheren Pflanzen. Berlin. Frey-Wyssling, A., 1935b. (Theory of junctions.) Proc. VI. Int. Bot. Congr. Amster- dam 2. 294. Illustrations to this topics in 1936a. Frey-Wyssling, a., 1935c. („Nutrition line".) Naturwiss. 23. 767. Frey-Wyssling, A., 1936a. (Structure of cell walls.) Protoplasma 25. 261. Frey-Wyssling, A., 1936b. (Cellulose, optical dispersion.) Helv. chim. Acta 19. 900. Frey-Wyssling, A., 1936c. (Filaments, optics.) Ber. dtsch. bot. Ges. 54. 445. Frey-Wyssling, A., i936d. (Cellulose fibres, permutoid reactions.) Protoplasma 26. 45. Frey-Wyssling, a., 1937a. (Intermicellar system. X-ray analysis.) Protoplasma 27. 372. Frey-Wyssling, A., 1937b. (Intermicellar system, ultramicroscopic analysis.) Proto- plasma 27. 563. Frey-Wyssling, A., 1937c. (Chloroplasts, structure.) Protoplasma 29. 279. Frey-Wyssling, A., i937d. (Submicroscopic morphology.) Ber. dtsch. bot. Ges. 55. (i 19). Frey-Wyssling, A., 1938. (Micellar science.) Kolloid-Z. 85. 148. Frey-Wyssling, A., 1940a. (Cytoplasm, fine-structure.) J. R. micr. Soc. 60. 128. Frey-Wyssling, a., 1940b. (Starch grains, optics.) Ber. schweiz. bot. Ges. 50. 321. Frey-Wyssling, A., 1940c. (Starch grains, optics.) Naturwiss. 28. 78. Frey-Wyssling, a., 1941. (Crossed micellar systems.) Protoplasma 35. 527. Frey-Wyssling, a., 1942. (Plant cell walls with tubular texture.) Jahrb. wiss. Bot.90. 705. Frey-Wyssling, a., 1943a. (Scattering in gels.) Helv. chim. Acta 26. 833. Frey-Wyssling, A., 1943b. (Chromosomes, scattering of nucleic acid chains.) Chromo- soma 2. 473. Frey-Wyssling, A., 1944a. (Chromosomes, distribution of nucleic acids.) Schweiz. med. Wochenschr. 74. 330. Frey-Wyssling, A., 1944b. (Genes, structure and size.) Arch. Klaus-Stift. 19. 451. Frey-Wyssling, a., 1945a. (Cell extension.) Arch. Klaus-Stift. 20. Erganzungsbd. p. 381. Frey-Wyssling, a., 1945b. Ernahrung und Stoffwechsel der Pflanzen. Zurich. Frey-Wyssling, A., 1947/49. (Plasma gel, protoplasmic flow.) Exp. Cell Research Suppl. I, Stockholm, p. 33. Frey-Wyssling, A., 1948a. (Extension growth, energetics.) Viertelj. schr. Naturf. Ges. Zurich 93. 24. Frey-Wyssling, a., 1948b. (Cell walls, growth in area.) Growth Symp. 12. 151. Frey-Wyssling, a., 1948c. (Starch grains, fine -structure.) Schweiz. Brauerei Rundschau 1948, No. 1. 382 LITERATURE Frey-Wyssling, a., 1949a. (Cytoplasm, physicochemical behaviour.) Research 2. 300. Frey-Wyssling, a., 1949b. (Chloroplasts, lipoproteins.) Faraday Soc. Disc. 1949, No. 6. p. 130. Frey-Wyssling, A., 195 1. Elektronenmikroskopie. Neujahrsblatt der Naturf. Ges. Zurich 195 1. Frey-Wyssling, A. and Frey, R., 1950. (Tunicin, electron microscopy.) Protoplasma 39, 656. Frey-Wyssling, a. and Muhlethaler, K., 1944. (Gels, electron microscopic studies.) Viertelj. schr. Naturf. Ges. Zurich 89. 214. Frey-Wyssling, a. and Muhlethaler, K., 1946. (Electron microscopy of bacterial cellulose.) J. Polymer Sci. i. 172. Frey-Wyssling, A. and Muhlethaler, K., 1949a. (Chloroplasts, electron microscopy.) Viertelj. schr. Naturf. Ges. Zurich 94. 179. Frey-Wyssling, A., and Muhlethaler, K., 1949b. (Root hairs, electron microscopy.) Mikroskopie (Wien) 4. 257. Frey-Wyssling, a. and Muhlethaler, K., 1949c. (Rayon, electron microscopy.) Schweiz. Bauzeitg. 67. 51. Frey-Wyssling, A. and Muhlethaler, K., 1950. (Chitinous cell walls, electron micros- copy.) Viertelj .schr. Naturf. Ges. Zurich 95. 45. Frey-Wyssling, A., Muhlethaler, K. and Wyckoff, R. W. G. 1948. (Cell walls, micro- fibrils.) Experientia 4. 475. Frey-Wyssling, a., and Speich, H., 1942. (Cellulose fibres, deficit of density.) Helv. chim. Acta 25. 1474. Frey-Wyssling, A.and Stecher, H., 1951. (Cell walls, growth in area.) Experientia 7. 420. Frey-Wyssling, A. and Steinmann, E., 1948. (Chloroplasts, layer birefringence.) Bio- chem. Biophys. Acta 2. 254. Frey-Wyssling, H. and Walchli, O., 1946. (Silver dichroism.) J. Polymer Sci. i. 266. Frey-Wyssling, A. and Weber, E., 1941. (Double refraction of flow, measurement.) Helv. chim. Acta. 24. 278. Fricke, H., 1925. (Erythrocytes, electrical capacity.) J. gen. Physiol. 9. 137. Friedel, G., 1922. (Mesophases.) Ann. Phys. Paris 18. 358. Friedrich-Freksa, H., Kratky, O. and Sekora, A., 1944. (Silk fibroin.) Naturwiss. 32. 78. Fritz, F., 1935. (Cuticular layers.) Jahrb. wiss. Bot. 81. 718. Fritz, F., 1937. (Rhythmical cutinisation.) Planta 26. 693. Gaffron, H. and Wohl, K., 1936. (Theory of COg assimilation.) Naturwiss. 24. 81, 103. Gaumann, E. and Jaag, O., 1936. (Cuticular transpiration.) Ber. schweiz. bot. Ges. 45. 411. Gause, G. F., 1936. (Chemical asymmetry.) Ergebn. Biol. 13. 54. Geitler, L., 1934. Grundriss der Zytologie. Berlin. Geitler, L., 1937. (Chloroplasts, grana.) Planta 26. 463. Geitler, L., 1938. Chromosomenbau. Berlin. Geitler, L., 1940. (Endomitosis.) Ber. dtsch. bot. Ges. 58. 131. Gerendas, M. and Matoltsy, A. G., 1947. (Striated muscle, optics.) Hung. Acta Physiol. I. No. 4. Gerngross, O., Herrmann, K. and Abitz, W., 1930. (Gelatin.) Biochem. Z. 228. 409. Gerngross, O., Herrmann, K. and Lindemann, R., 1932. (Gelatin.) KoUoid-Z. 60. 276. Gerngross, O. and Katz, J. R., 1926. (Gelatin, X-ray diffraction.) Kolloid-Z. 39. 181. Gibbons, G. C. and Boissonnas, R. A., 1950. (Amylopectine, branching 1-6.) Helv. chim. Acta 33. 1477. Gicklhorn, J., 1932a. (Myelin forms.) Protoplasma 15. 90. LITERATURE 585 GiCKLHORN, J., 1932b. (Nucleus, protein crystals.) Protoplasma 15. 276. GiROUD, A., 1928. (Mitochondria, optics.) C.r. Acad. Sci. Paris 186. 794. GoLDACRE, R. J. and Lorch, I. J., 1950. (Plasm flow.) Nature 166. 497. GoLDSCHMiDT, V. M., 1926. (Diameter of atoms.) Naturwiss. 14. 477. GoLDSCHMiDT, V. M., 1927. (Diameter of atoms.) Ber. dtsch. chem. Ges. 60. 1263. Gorter, E. and Grendel, F., 1925. (Erythrocytes, lipid coating.) J. exp. Med. 41. 439. GoRTER, E. and van Ormondt, J., 1935. (Myosin, spreading.) Biochem. J. 29. 48. Gorter, E., van Ormondt, J. and Meyer, T. M., 1935. (Complex proteins, spreading.) Biochem. J. 29. 38. GouGH, A., 1924. (Erythrocytes, shape.) Biochem. J. i8. 202. Graham, Th., 1861. (Colloid particles.) Philos. Trans. 1861, p. 183. Graham, Th., 1862. (Dialysis.) Liebigs Ann. i2i. i. Grakick, S., 1938. (Chloroplasts, method of isolation.) Amer. J. Bot. 25. 561. Granick, S., 1948. (Haem and chlorophyll.) Harvey Lectures (Springfield, 111.) 44. 220. Granick, S., 1949. (Chloroplast, number of grana.) In Franck and Loomis, Photosyn- thesis in plants. Ames, Iowa, p. 113. Granick, S. and Porter, K. R., 1947. (Chloroplasts, electron microscopy.) Amer. J. Bot. 34- 545- Gross, J., 1949. (Elastic tissue, electron microscopy.) J. exp. Med. 89. 699. Gross, J., 1950. (Collagen fibres, segmented.) Ann. N.Y. Acad. Sci. 52. 964. Guilliermond, a., 1930. (Protoplasm, ultramicroscopy.) Rev. gen. Bot. 42. 327. GuiLLiERMOND, A., Mangenot ,G. and Plantefol,L. 1933. Traite de Cytologic vegetale, Paris. GuNDERMANN, J., Wergin, W. and Hess, K., 1937. (Waxes of cell walls.) Ber. dtsch. chem. Ges. 70. 517. Haas, J. N., 1950. (Muscle fibre, tip growth.) Growth 14. 277. Haase-Bessel, G., 1936. (Substratum of genes.) Planta 25. 240. Hadorn, E., 1939. (Pupation, caused by hormones.) Mitt. Naturwiss. Ges. Thun 1939, Heft 4, p. I. Hakansson, a. and Levan, A., 1942. (Nucleoli, formation.) Hereditas (Lund) 28. 436. Hall, C. E., Jakus, M. A. and Schmitt, F. O., 1945. (Muscle, electron stains.) J. appL Phys. 16. 459. Hall, C. E., Jakus, M. A. and Schmitt, F. O., 1946. (Striped muscle fibres, electron micrograph.) Biol. Bull. 90. 32. Halle, F., 193 i. (Length of rodshaped molecules.) Kolloid-Z. 56. 77. Halle, F., 1937. (Protein molecules, structure.) Kolloid-Z. 81. 334. Hanes, C. S., 1937. (Starch, hydrolysis.) New Phytol. 36. loi. Hanes, C. S., 1940. (Synthesis of starch.) Proc. Roy. Soc. London 128. 421, 129. 174. Hanson, E. A. and Katz, J. R., 1934. (Starch grains.) Z. phys. Chem. A 168. 339. Hansteen-Cranner, B., 1926. (Phosphatides in the cell wall.) Planta 2. 438. Harder, R., 1937. (Chitin of fungi.) Nachr. Ges. Wiss. Gottingen (math. -phys. Kl.) Nachr. Biol. 3. i. Harris, J. E., 1939. (Nuclear thixotropy.) J. exp. Biol. 16. 258. Harris, J. E., 1952. (Nucleus, thixotropy.) In A. Frey-Wyssling, Deformation and flow in biological systems. Amsterdam, p. 486. Harvey, E. B., 1933. (Cell fragments, development.) Biol. Bull. 64. 125. Harvey, E. B., 1936. (Centrifuged cells.) Biol. Bull. 71. loi. Harvey, E. N., 1936. (Surface tension, sessile drop.) J. cell. comp. Physiol. 8. 251. Harvey, E. N., 1937. (Cytoplasm, surface tension.) Trans. Faraday Soc. 33. 943. Harvey, E. N. and Loomis, A. L., 1930. (Centrifuging microscope.) Science 72. 42. 384 LITERATURE Harvey, E. N. and Danielli, J. F., 1936. (Films of proteins, elasticity.) J. cell. comp. Physiol. 8. 31. Haurowitz, p., 1949. (Protein molecules, internal structure.) Experientia 5. 347. Hausermann, E., 1944. (Intercellular spaces, wettability.) Diss. E. T. H. Ziirich. Haworth, W. N., 1925. (Sugars, pyranose structure.) Nature 116. 430. Haworth, W. N., 1929. The constitution of sugars. London. Hecht, K., 1912. (Plasmolysis.) Beitr. Biol. Pfl. ii. 137. Hegetschweiler, R., 1948. (Silk, double refraction.) Viertelj.schr. Naturf. Ges. Ziirich 93- 57- Hegetschweiler, R., 1950. (Silk fibroin, optics.) Diss. E. T. H. Zurich; Makromol. Chem. 4. 156. Heidenreich, L. D. and Matheson, L. A., 1944. (Surface orientation, electron microsco- pic studies.) J. appl. Phys. 15. 423. Heierle, E., 1935. (Pigments of leaves, spectrometric analysis.) Diss. E.T.H. Zurich. Heilbronn, a., 1914. (Cytoplasm, viscosity.) Jahrb. wiss. Bot. 54. 357. Heilbronn, a., 1922. (Cytoplasm, displacement of magnetic particles.) Jahrb. wiss. Bot. 61. 284. Heilbrunn, L. v., 1930. (Cytoplasm, viscosity.) Protoplasma 8. 58. Heitz, E., 193 1. (Nucleoli, number.) Planta 12. 775. Heitz, E. 1932. (Chloroplasts, grana.) Planta 18. 616. Heitz, E., 1935. (Chromosomes, structure.) Z. indukt. Abstamm.- u. Vererblehreyo. 402. Heitz, E., 1936a. (Chloroplasts, structure.) Ber. dtsch. bot. 54. 362. Heitz, E., 1936b. (Chloroplasts, structure.) Planta 26. 134. Hengstenberg, J., 1928. (Intramolecular periods.) Z. Kristallogr. 67. 583. Hengstenberg, J. and Mark, H., 1928. (X-ray determination of particle size.) Z. Kristal- logr. 69. 271. Hermans, P. H., 1938. (Deformation of gels.) Kolloid-Z. 83. 71. Hermans, P. H., 1941. (Swelling of gels.) Kolloid-Z. 97. 231. Hermans, P. H., 1946. Contribution to the physics of cellulose fibres. Monograph on the progress of research in Holland during wartime. Amsterdam. Hermans, P. H., 1949. Physics and Chemistry of Cellulose Fibres. New York and Amsterdam. Hermans, P. H. and De Leeuw, A. J., 1937, 1941. (Isotropic cellulose fibres.) Kolloid-Z. 8i. 300, 97. 326. Hermans, P. H. , Kratky, O. and Treer, R., 1941. (Gels, orientation during extension.) KoUoid-Z. 96. 30. Hermans, P. H. and Vermaas, D., 1946. (Cellulose fibres, density.) J. Polymer Sci. i. 149. Hermans, P. H. and Weidinger, A., 1949. (Crystallinity of cellulose.) J. Polymer Sci. 4. 135. Herzog, R. O., 1927. (Cellulose „crystallits".) Handb. d. Textilfasern 7. i. Berlin. Hess, K., Kiessig, H. and Gundermann, J., 1941. (Cellulose crushed in ball mill.) Z. phys. Chem. B. 49. 64. Hess, K., Kiessig, H., Wergin, W. and Engel, W., 1939. (Cell wall, formation.) Ber. dtsch. chem. Ges. 72. 642. Hess, K., Trogus, C. and Wergin, W., 1936. (Meristems, „Primarsubstanz".) Planta 25. 419. Heyn, a. N. J., 1931. (Cell elongation, optics.) Diss. Utrecht. Heyn, a. N. J., 1933. (Epidermis, X-ray analysis.) Proc. Acad. Sci. Amsterdam 36. 560. Heyn, A. N. J., 1934. (Meristems, X-ray analysis.) Protoplasma 21. 299. Hill, A. V., 1950. (Muscle, mechanics of contraction.) Nature 166. 415. HiLLiER, J., 1946. (Electron microscopy.) Bull. Amer. Ceramic Soc. 25. 438. LITERATURE 385 HiLLiER, |. and Gettner, M. E., 1950. (Electron microscopy, thin sections.) Science 112. 520. HiRSCHLER, J., 1942. (Chromosomes, contents of lipids.) Naturwiss. 30. 642. HoAGLAND, D. R. and Broyer, T. C, 1936. (Resorption and respiration.) Ann. Bot. Lond. 50. 345. HoBER, R., 1922. Physikalische Chemie der Zelle und der Gewebe, Leipzig. Hoerr, N., 1943. (Mitrochondria, isolation.) Biol. Symp. 10. 185. HoFLER, K., 193 1. (Plasmic membranes.) Ber. dtsch. bot. Ges. 49. (79). HoFLER, K., 1932. (Tonoplast.) Protoplasma 15. 462. HoFLER, K., 1934. ("Intrability".) Z. wiss. Mikr. Kiister-Festschr. 51. 70. HoFLER, K., 1936a. (Compared permeability.) Mikrochem. Molisch-Festschr. p. 224. HoFLER, K., 1936b. (Rhodophyceae, centrifuging experiments.) Protoplasma 26. 377. HoFLER, K., 1940. (Cap plasmolysis.) Ber. dtsch. bot. Ges. 58. 292. HoFLER, K., 1942. (Specific permeability series.) Ber. dtsch. bot. Ges. 60. 179. HoFMEiSTER, L., 1935. (Compared premeability.) Bibl. bot. Stuttgart 113. i. HoFMEiSTER, L., 1940. (Plastids of diatoms.) Z. wiss. Mikr. 57. 259. HoGEBOOM, G. H., Claude, A. and Hotchkiss, R. D., 1946. (Mitochondria, fixed oxida- ses.) J. biol. Chem. 165. 615. HoHNEL, F. VON, 1887. Mikroskopie der Faserstoffe. Leipzig. HoiwiNK, A. L., 1951. (Flagellum of Euglena gracilis.) Proc. Acad. Sci. Amsterdam C 54. 132. HuBER, L. and Rouiller, Ch., 195 i. (Collagen fibre of bone.) Experientia 7. 338. Hubert, B., 1955. (Chloroplasts, molecular structure.) Diss. Leiden. HiJRTHLE, K., 193 1. (Muscle fibre, fine-structure.) Pfiiig. Arch. ges. Physiol. 227. 610. HusEMANN, E. and Ruska, H., 1940. (Molecules of glycogen.) J. prakt. Chem. 156. i. Naturu'iss. 28. 534. HusKiNS, C. L., 1941. (Chromonema, spiral structure.) Cold Spr. Harb. Symp. quant. Biol. 9. 13. HusKiNS, C. L., 1942. (Nucleus, fine-structure.) In W. Seifriz, The structure of protoplasm. Ames-Iowa, p. 109. HusTED, F., 1930. Bacillariophyta (Diatomeae.) 2. Aufl. Jena. Huxley, H. E. and Perutz, M. F., 1951. (Muscle fibre. X-ray analysis.) Nature 167. 1054. Induni, G., 1945. (Swiss electron microscope.) Viertelj. schr. Naturf. Ges. Zurich 90. 181. Irvine, J. C, 1932. (Starch methylated.) Nature (Lond.) 129. 470. Itschner, v., 1935. (Glucosamin, constitution.) Diss. Univ. Zurich. Jaccard, p. and Frey, A. 1928. (Tracheids, spiral texture.) Jahrb. wiss. Bot. 68. 844. Jakus, M. a., 1945. (Trichocysts of Paramecium, electron micrograph.) J. exp. Zool. 100. 457. Jaloveczky, p., 1942. (Starch grains, lamellar structure.) Bot. Kozlen (Ungarn) 39. 610. Jeener, R., 1948. (Cytoplasm, submicroscopic granules.) Biochim. Biophys. Acta 2. 635. Jensen, O. S.,.1887. (Spermatozoa, spiral body.) Arch. mikr. Anat. 30. Joly, M., 1948. (Proteins, spreading.) Biochim. Biophys. Acta 2. 624. Joly, M. 1949. (Beaded protein chains.) Kolloid-Z. 115. 83. Jordan, P., 1947. Eiweissmolekiile. Stuttgart. JoRPES, E., 1932. (Erythrocytes, stromatin.) Biochem. J. 36. 1488. Jung, F., 1942. (Erythrocytes, fine-structure.) Klin. Wochenschr. 21. 917. Jung, F., 1950. (Erythrocyte, review on structure.) Naturwiss. 37. 229, 254. Kamiya, N., 1940 (Protoplasmic flow, propulsive pressure.) Science 92. 462. Kamiya, N., 1942. (Protoplasmic flow, propulsive pressure.) In W. Seifriz, The structure of protoplasm. Ames-Iowa, p. 199. Kamiya, N. and Abe, S., 1950. (Myxomycetes, plasm flow.) J. Colloid Sci. 5. 149. 386 LITERATURE Karrer, p., 1935. (Carotenoids, constitution.) Schweiz. med. Wochenschr. 65. 898. Karrer, p., 1941. Lehrbuch der organischen Cliemie. 7. Aufl. Leipzig. Karrer, P., 1944. (Vitamins.) Schweiz. Z. Pathol. Baicteriol. 7. 303. Karrer, P., 1946. (Carotenoids, epioxides.) Bull. Soc. Chim. bid. 28. 688. Karrer, P. Peyer, I. und Zegar, Z., 1923. (Cellulose in cork.) Helv. chim. Acta 5.853. Katz, J. R., 1924. (Swelling.) Ergebn. exakt. Naturwiss. 3. 332. Katz, J. R. and Derksen, J. C, 1933. (Starch, X-ray diffraction.) Z. phys. Chem. A 165. 228. Katz, J. R. and De Rooy, A., 1933. (Blood fibrin.) Naturwiss. 2X. 559. Kausche, G. A. and Ruska, H., 1940. (Chloroplasts, electron micrograph.) Naturwiss. 28. 303. Kelaney, M. a. El and Searle, G. O., 1930. (Cellulose fibres, chemical cross-sections.) Proc. Roy. Soc. London 106. 357. Keller, R., 1932. Die Elektrizitat der Zelle. 3. Aufl. Mahrisch Ostrau. Kerr, Th., 1937. (Cotton hairs, growth layers.) Protoplasma 27. 229. KiESEL, A., 1930. Chemie des Protoplasmas. Berlin. Kinsinger, W. G., Hillier, J., Picard, R. G. and Zieler, H. W., 1946. (Electron microscope, resolution.) J. appl. Phys. 17. 989. Knapp, a., 1946. (Nucleic acid, double refraction of flow.) Diss. Bern. Knoch, M. and Konig, H., 1951. (Trichocysts of Paramecium.) Naturwiss. 38. 531. Koepfli, J. B., Thimann, K. V. and Went, F. W., 1938. (Auxins, chemistry.) J. biol. Chem. 122. 763. Kohler, a., 1904. (UV microscopy.) Z. wiss. Alikr. 21. 129, 273. KossEL, A., 1905. (Protamins.) Hoppe Seyler Z. physiol. Chem. 44. 347. KossEL, A., 1929. Protamine und Histone. Wien. Kratky, O., 1933. (Orientation in gels.) Kolloid-Z. 64. 213. Kratky, O., 1934. (Short-range order.) Kolloid-Z. 68. 347. Kratky, O., 1938. (Orientation in gels.) Kolloid-Z. 84. 149. Kratky, O., 1940. (Cellulose gels, fine-structure.) Angew. Chem. 53. 153. Kratky, O. and Kuriyama, S., 1931. (Silk fibroin.) Z. phys. Chem. B 11. 363. Kratky, O. and Mark, H., 1937. (Cellulose micelles as individuals.) Z. phys. Chem. B 36. 129. Kratky, O. and Platzek, P., 1938. (Intrinsic double refraction, depending on swelling.) Kolloid-Z. 84. 268. Kratky, O., Schauenstein, E. and Sekora, A., 1950. (Silk I and silk IL) Nature 165. 319. Kratky, O. and Schneidmesser, B., 1938. (Schardinger dextrin. X-ray analysis.) Ber. dtsch. chem. Ges. 71. 141 3. Kreger, D., 1946. (Starch grain. X-ray diffraction.) Nature 158. 199. Kreger, D. R., 1949. (Plant waxes, X-ray study.) Rec. Trav. bot. neerl. 41. 603. Kreger, D. R., 195 i. (Starch grains. X-ray analysis.) Biochim. Biophys. Acta 6. 407. KuHN, R., 1937. (Carotenes.) Naturwiss. 25. 225. KuHN, W. and Hargitay, B., 1951. (Chain molecules, contraction.) Experientia 7. i. KiJNTZEL, A., 1929. (Tendon, rodlet double refraction.) Collegium 4. 207. KuNTZEL, A., 1941. (Collagen fibre.) Kolloid-Z. 96. 273. KuNTZEL, A. and Prakke, F., 1933. (Collagen, swelling.) Biochem. Z. 267. 243. KiJSTER, E., 1933. (Plastids, anisotropy.) Ber. dtsch. bot. Ges. 51. 523. KiJSTER, E., 1934a. (Protein spindles.) Ber. dtsch. bot. Ges. 52. 564. KusTER, E., 1934b. (Nucleus, anisotropy.) Ber. dtsch. bot. Ges. 52. 626. KiJSTER, E., 1935a. Die Pflanzenzelle. Jena. KiJSTER, E., 1935b. (Plastids, anisotropy.) J. R. inicr. Soc. 40. 99. LITERATURE 587 KiJSTER, E., 1935c. (Plastids, formation of strands.) Ber. dtsch. bot. Ges. 53. 834. KiJSTER, E., 1937. (Plastids, anisotropy.) Z. wiss. Milcr. 54. 88. KuwAD.\, Y. and Nak.\ml'r.\, T., 1934. (Chromonema.) Cytologia 6. 78. Lagerm.\lm, G., Philip, B. and Lindberg, J., 195 1. (Finger nail, surface film.) Nature 168. 1080. Lamm, O., 1957. (Potato starch.) Nova Acta Soc. Sci. upsal. 10. No. 6. Landolt-Bornstein, 1923. Physikalisch-chemische Tabellen. Berlin. Langelaan, J. W., 1946. (Muscle fibre, interference pattern.) Verb. Akad. Wet. Amster- dam, 2. Sect. 42. No. 4, p. I. Langmuir, J., 1917. (Surface films.) J. Amer. chem. Soc. 39. 1848. Laue, M. von, 1926. (Particle size by X-ray analysis.) Z. Kristallogr. 64. 115. Lea, D. A., 1940. (Size of genes as determined by ionizing rays.) Nature 146. 137. Lecher, E., 1919. Lehrbuch der Physik Rir Mediziner. Leipzig. Lehmai^tn, F. E., 1937a. (Disturbance in development by LiCl.) Naturwiss. 25. 124. Lehma>jn, F. E., 1937b. (Disturbance in development by LiCl.) Arch. Entw. Mech. Org. 136. 112. LEHM.A>rN, F. E., 1940. (Tubifex eggs, polarity.) Rev. Suisse Zool. 47. 177. Lehmann, F. E., 1945. Einfuhrung in die physiologische Embryologie. Basel. Lehm.ann, F. E., 1947. (Biosomes.) Rev. Suisse Zool. 54. 246. Lehmann, F. E., 1948. (Egg of Tubifex, polar plasm.) Rev. Suisse Zool. 55. i. Lehmann, F. E., 1950. (Cytoplasm, submicroscopic globular particles.) Experientia 6. 382 Lehmann, F. E., 195 i. (Beaded chains in Amoeba.) Kolloqu. dtsch. Ges. physiol. Chem. Mosbach-Baden, p. i. Lehmann, F. E. and Biss, R., 1949. (Tubifex egg, electron microscopy.) Rev. Suisse Zool. 56. 264. Lehmann, O., 1917. (Liquid crystals.) Ergebn. Physiol. i6. 255. Lepeschkin, W. W., 1927. (Er}'throcyte, membrane.) Protoplasma 2. 143. Leuthardt, F., 1949. (Mitochondria, metaboHsm.) Viertelj.schr. Naturf. Ges. Zurich 94. 132. Leuthold, p., 1933. (Aleurone grains.) Ber. schweiz. bot. Ges. 42. 31. Lewis, E. B., 1950. (Genes, screened by heterochromatin.) Adv. in Genetics 3. 73. Lewis, W. H., 1942. (Plasm viscosity.) In W. Seifriz, The structure of protoplasm. Ames-Iowa, p. 163. Liebich, H., 1941. (Chloroplasts, contents of iron.) Z. Bot. 37. 129. Lieser, Th., 1940. (Cellulose, supramolecular constitution.) Cellulose-Chem. 18. 121. LiESER, Th. et al., 1941. (Cellulose, supramolecular constitution.) Liebigs Ann. 548, 195, 204, 212. Lindberg, J., Philip, B. and Gralen, N.,1948. (Wool, epidermicula.) Nature 162. 458. LiNDLEY, H., 1950. (Wool structure, review.) Research 3. 509. LoEB, J. and Beutner, R., 1912/13. (Potential in membranes.) Biochem. Z. 41. i, 44. 303, 51. 300. LoEWY, A. G., 1949. (Protoplasmic flow.) Amer. Philosoph. Soc. 93. 326. LoTMAR, W. and Picken, L. E. R., 1950. (Chitin of Polychaeta and Mollusca.) Exper- ientia 6. 58. LiJDTKE, M., 1936. (Cellulose fibres, "Querelemente".) Planta 25. 791. Lundegardh, H., 1922. Zclle und Zytoplasma. Berlin. LuNDEGARDH, H. and Burstrom, H., 1933. (Resorption and respiration.) Planta 18. 683. LiiscHER, E., 1936. (Cork waxes.) Diss. Bern. LuYET, B. J., 1937. (\^itrification of gels.) Biodynamica 1937, No. 29. LuYET, B. J., 1939. (Cytoplasm, freezing.) Arch. exp. Zellforsch. 22. 487. 388 LITERATURE LuYET, B. J. and Ernst, R., 1934a. (Nuclear frame and nuclear sap.) Biodynamica 1934. No. 2. LuYET, B. J. and Ernst, R., 1934b. (Nuclear membrane.) Proc. Soc. exp. Biol. 31. 1225. McArthur, I., 1943. (Wool, X-ray analysis.) Nature 152. 38. McClintoc, B., 1950. (Genes, influence of heterochromatin.) Proc. nat. Acad. Sci. Wash. 36. 344. Mackinney, G. and Milner, H. W., 1933. (Carotenes.) J. Amer. chem. Soc. 55. 4728. Magnus, A., 1922. (Numbers of coordination.) Z. anorg. Chem. 124. 291. Manegold, E., 1941. (Systematics of gels.) KoUoid-Z. 96. 186. Manton, I. and Clarke, B., 1950. (Spermatozoid of Fucus, electron microscopy.) Nature 166. 973. Mark, H., 1940. AUgemeine Grundlagen der hochpolymeren Chemie. Leipzig. Mark, H. and Philipp, H., 1937. (Proteins, crystal lattice.) Naturwiss. 25. 119. Mark, H. and Schossberger, F., 1937. (Organic compounds, crystal structure.) Ergebn. exakt. Naturwiss. 16. 183. Marklund, G., 1936. (Compared permeabiHty.) Acta hot. fenn. 18. Marks, M. H., Bear, R. S. and Blake, C. H., 1949. (Collagen family.) J. exp. Zool. III. 55. Marsland, D. a., 1942. (Cytoplasm, high pressure experiments.) In W. Seifriz, The structure of protoplasm. Ames-Iowa, p. 127. Martens, P., 1927/29. (Karyokinesis, vital observation.) Cellule 38. 67, 39. 167. Martens, P., 1934. (Cuticle, origin.) Cellule 43. 289. Martin, L. C, 1938. (Electron microscope.) Nature 142. 1062. Matoltsy, a. G. and Gerendas, M., 1947. (Striated muscle, optics.) Nature 159. 502. Matoltsy, A. G., Gross, J. and Grignolo, A., 1951. (Vitreous body of eye, beaded chains.) Proc. Soc. exp. Biol. 76. 857. Matthey, R., 1941. (Chromosomes, spiral structure.) Arch. Klaus-Stift. 16. 630. Mazia, D. and Jaeger, L., 1939. (Chromosomes, digestion by nuclease.) Proc. nat. Acad. Sci. Wash. 25. 456. Meeuse, a. D. J., 1938. (Sclerenchyma fibres, growth.) Rec. Trav. bot. neerl. 35. 288. Meeuse, a. D. J., 1941. (Sliding growth.) Diss. Delft. Menke, W., 1934a. (Chloroplasts, myelin forms.) Protoplasma 21. 279. Menke, W., 1934b. (Chloroplasts, double refraction.) Protoplasma 22. 56. Menke, W., 1938a. (Cytoplasm, chemistry.) Z. Bot. 32. 273. Menke, W., 1938b. (Chloroplasts, method of isolation.) Hoppe Seyler Z. physiol. Chem. 257- 43- Menke, W., 1938c. (Chloroplasts, fine-structure.) Kolloid-Z. 85. 256. Menke, W., 1940a. (Chloroplasts, electron micograph.) Protoplasma 35. 115. Menke, W., 1940b. (Chloroplasts, contents of chlorophyll and ash.) Hoppe Seyler Z. physiol. Chem. 263. 100, 104. Menke, W., 1940c. (Carotenoids of plastids as chromoproteins.) Naturwiss. 28. 31. Menke, W., i94od. (Chloroplasts, lamination.) Naturwiss. 28. 158. Menke, W., 1943. (Plastids, dichroism.) Biol. Zbl. 63. 326. Menke, W., and Jacob, E., 1942. (Chloroplasts, contents of lipids.) Hoppe Seyler Z. physiol. Chem. 272. 227. Menke, W. and Koydl, E., 1939. (Chloroplasts, lamination.) Naturwiss. 27. 29. Menke, W. and Kuster, H. J., 1938. (Chloroplasts, dichroism.) Protoplasma 30. 283. Mercer, E. H., 1951. (Fibroinogen of silk.) Nature 168. 792. Mestre, H., 1930. (Phyllochlorin.) Contr. Marine Biol. Stanford Univ. Press. Metz, C. W., 1941. (Salivary gland chromosomes, structure.) Cold Spr. Harb. Symp. quant. Biol. 9. 23. LITERATURE 3^9 Metzner, p., 1937. (Chloroplasts, fluorescence.) Ber. dtsch. bot. Ges. 55. (16). Meyer, A., 1883. Das Chlorophyllkom. Leipzig. Meyer, A., 1895. Untersuchungen iiber die Starkelcorner. Jena. Meyer, K. H., 1929. (Muscle fibre, contractibility.) Biochem. Z. 214. 253. Meyer, K. H., 1930. (Model of micelles.) Kolloid-Z. 53. 8. Meyer, K. H., 1935. (Theory of selective permeability.) Lecture 29th Nov. Ass. des Chimistes de Geneve. Meyer, K. H., 1940a. Die hochpolymeren Verbindungen. Leipzig. Meyer, K. H., 1940b. (Ramified and unramified starch components.) Naturwiss. 28. 397- Meyer, K. H., 1943. (Starch, review.) Melliand Textilbcr. 1943, No. 3. Meyer, K. H., 1951. (Amylases.) Angew. Chem. 63. 153. Meyer, K. H. and Bernfeld, P., 1941a. (Starch, blue iodine reaction.) Helv. chim. Acta 24. 389. Meyer, K. H. and Bernfeld, P., 1941b. (Potato starch). Helv. chim. Acta 24. 378. Meyer, K. H. and Bernfeld, P., 1941c. (Action of a-amylase.) Helv. chim. Acta Fest- schr. Engi 1941, p. 359. Meyer, K. H. and Bernfeld, P., 1946. (Membrane structure, potentiometric analysis.) J. gen. Physiol. 29. 353. Meyer, K. H. and Brentano, W., 1936. (Corn starch.) C. R. Soc. Phys. Hist. nat. Geneve 53. iii. Meyer, K. H., Fischer, E. H. and Bernfeld, P., 1947. (a-amylase, crystallisation.) Ex- perientia 3. 106, 455. Meyer, K. H., Fischer, E. H. and Piguet, A., 195 1. (/S-amylase, crystallisation.) Helv. chim. Acta 34. 316. Meyer, K. H. and Heinrich, P., 1942. (Starch, content of amylopectin.) Helv. chim. Acta 25. 1639. Meyer, K. H. and Mark, H., 1928. (Silk fibroin as main valency chain lattice.) Ber. dtsch. chem. Ges. 61. 1932. Meyer, K. H. and Mark, H., 1930. Der Aufbau der hochpolymeren organischen Natur- stofFe. Leipzig. Meyer, K. H. and Mark, H., 1940. Hochpolymere Chemie. 2 Bde. Leipzig. (See Mark, H., 1940, Meyer, K. H., 1940a). Meyer, K. H. and Misch, L., 1937. (Cellulose, crystal lattice.) Helv. chim. Acta 20. 232. Meyer, K. H. and Pankow, G. W., 1935. (Chitin, crystal lattice.) Helv. chim. Acta 18. 589. Meyer, K. H. and Picken, L. E. R., 1937. (Muscle fibre, thermodynamics.) Proc. Roy. Soc. London B 124. 29. Meyer, K. H. and Sievers, J. F., 1936. (Theory of selective permeability.) Helv. chim. Acta 19. 649, 665, 948, 987. Meyer, K. P., 1939. (Chlorophyll, spectroscopy.) Helv. phys. Acta 12. 349. Meyer, M., 1938. (Cuticular layers, submicroscopic structure.) Diss. E.T.H. Zurich. MiCHAELis, L., 1933. Oxydations- und Reduktionspotentiale. Berlin. Michel, K., 1940. Grundziigc der Mikrophotographie. Jena. Michener, H. D., 1935. (Ethylene, stimulating substance.) Science 82. 551. Middlebrook, W. R., 195 1. (Wool keratin, chemical composition.) Biochim. Biophys. Acta 7. 547- Minder, W. and Liechti, A., 1945. (Irradiation, effects in biology.) Experientia i. 298. MiRSKY, A. E., 1936. (Formation of protein frame after fertihzation.) Science 84. 333. MiRSKY, A. E. and Pollister, A. W., 1943. (Fibrous nucleoproteins.) Biol. Symp. 10. 247. MiTCHisoN, J. M., 1950a. (Amoeba, birefringence.) Nature 166. 313. Mitchison, J. AL, 1950b. (Erythrocyte, membrane.) Nature 166. 347. 590 LITERATURE MiTTASCH, A., 1936. (Catalysis.) Natunviss. 24. 770. MoHRiNG, A., 1922. (Chitin, rodlet double refraction.) Diss. Jena. Summary in Ambronn- Festschr. 1926, Kolloidchem. Beih. 23. 162. MoLLENDORF, W. V., 1937. (Cytoplasm, viscosity.) Arch. exp. Zellforsch. 19. 263. MoMMAERTS, W. F., 1938. (CHloroplasts, isolation of grana.) Proc. Acad. Sci. Amster- dam 41. 896, MoNNE, L., 1942a. (Cytoplasm, submicroscopic structure.) Ark. Zool. 34B. No. 5. MoNNE, L., 1942b. (Mitochondria, vital staining.) Arch. exp. Zellforsch. 24. 373. MoNNE, L., 1942c. (Nuclear membrane, double refraction.) Ark. Zool. 34B. No. 2. MoNNE, L., 1944. (Egg, cytoplasmic structure.) Ark. f. Zool. (Stockholm) 35A. Nr. 13. MoNNE, L., 1946a. (Chromidia.) Expericntia 2. 153. MoNNE, L., 1946b. (Sea urchin egg, polar organization.) Ark. Zool. 38 A No. 15. MoNNE, L., 1948. (Cytoplasm, chromidia.) Advances in Enzymol. 8. i. MoNROY, A., 1945. (Egg, cortical layer.) Experientia i. 335. MoNROY, A. and A., 1946. (Sea urchin egg, cortex.) Pubbl. Staz. zool. Napoli 20. 46. MoNROY, A. and Montalenti, G., 1947. (Egg, variation of double refraction.) Biol. Bull. (U.S.A.) 92. 131. Morton, T. H., 1935. (Dyes, particle size.) Trans. Faraday Soc. 31. I. 262. Mothes, K., 1933. (Dehydrogenation.) Fortschr. Bot. 2. 196. MoYER, L. S., 1942. (Protein and plasmic structure.) In W. Seifriz, The structure of protoplasm. Ames-Iowa, p. 23. Muhlethaler, K., 1949. (Bacterial cellulose, twined microfibrils.) Biochim. Biophys. Acta 3. 527. Muhlethaler, K., 1950a. (Developing cell walls, electron microscopy.) Biochim. Bio- phys. Acta 5. I. Muhlethaler, K., 1950b. (Cell walls, bipolar growth.) Ber. schweiz. bot. Ges. 60. 614. Muhlethaler, K., Muller, A. F. and Zollinger, H. U., 1950. (Mitochondria, membrane.) Experientia 6. 16. Muller, H. J., 1935. (Gene, size.) Amer. Nat. 69. 405. MiJLLER, A., 1929. (Paraffin chains.) Z. Kristallogr. 70. 386. MiJLLER, A. F. and Leuthardt, F., 1950. (Mitochondria, tricarbon acid cycle.) Helv. chim. Acta 33. 268. Muller, H. O., 1942a. (Electron microscope, focal depth.) Diss. Berlin. Muller, H. O., 1942b. (Electron microscopy, thickness of specimen.) KoUoid-Z. 99. 6. MiJLLER, H. O. and Pasewaldt, C. W. A., 1942. (Pleurosigma, electron micrograph.) Naturwiss. 30. 55. MuRALT, A. VON, 1932. (Muscle fibre, double refraction depending on contraction.) Pfiiig. Arch. 230. 299. MuRALT, A. VON, 1933. (Muscle fibre, fine-structure.) KoUoid-Z. 63. 228. MuRALT, A. VON, 1946. Die Signaliibermittlung im Nerven. Birkhiiuser, Basel. MuRALT, A. VON and Edsall, J. T., 1930. (Myosin, double refraction of flow.) Trans. Faraday Soc. 26. 837. Myrback, K., 1938. (Starch and dextrins.) Biochem. Z. 297. 160, 172, 179. Myrback, K., Ortenblad, B. and Ahlborg, K., 1942. (Starch and dextrins.) Biochem. Z. 311. 227; 315. 240. Nageli, C, 1858. Die Starkekorner. Zurich. Nageli, C, 1862. (Protein crystalloids.) S.B. bayer. Akad. Wiss. 4 (i) 290; 4. (2). 120. Nageli, C, 1928. Micellartheorie. Ostwalds Klassiker No. 227, herausgeg. von A. Frey. Leipzig. Nageli, C. and Schwendener, S., 1877. Das Mikroskop. Leipzig. Nageotte, J., 1936. (Myelin forms.) Actual, sci. industr. No. 431-434. Paris. LITERATURE 391 Nakamura, T., 1937. (Anisotropic chromosomes.) Cytologia, Tokyo, Fujii Jubilee Vol. p. 482. Naray-Szabo, St. v., 1928. (Starch, cr>'Stal lattice.) Liebigs Ann. 465. 299. Nathanson, a., 1904. (Permeability, mosaic theor}'.) Jahrb. wiss. Bot. 39. 607. Nebel, B. R., 1939. (Chromosomes, structure.) Bot. Rev. 5. 563. Nebel, B. R., 1941. (Chromonemata, number per chromosome.) Cold Spr. Harb. Symp. quant. Biol. 9. 7. Nebel, B. R. and Ruttle, M. L., 1937. (Chromonemata, number per chromosome.) Z. Zellforsch. 26. 281. Needham, D. M., 1942a. (Muscle contraction.) Ann. Rep. Progr. Chem. for 1941, p. 241. Needham, D. M., 1942b. (Adenosine triphosphatase.) Biochem. J. 36. 113. Needham, J. and D. M., 1925. (Redox potential.) C. R. Soc. Biol. Paris 93. 503. Nemec, B., 1929. (Nucleus, microscopic structure.) Protoplasma 7. 423. Newman, S.B., BoRYSKO, E. and Swerdlow, M., 1949. (Electron microscopy, thin sections.) Science no. 66. NicoLAi.E.and Frey-Wyssling,A., 1938. (CellwallofChaetomorpha.) Protoplasma 30.401. NicoLAi, M. F. E., and Weurman, C, 1938. (Chlorophyll multifilms.) Proc. Acad. Sci. Amsterdam 41. 904. NiGGLi, p., 1929. Geometrische Kristallographie des Diskontinuums. Berlin. NiGGLi, P., 1941/42. Lehrbuch der Mineralogie. 3. Aufl. Berlin. Noack, K., 1927. (Chlorophyll, state in chloroplast.) Biochem. Z. 183. 135. Noack, K., 1930. (Chloroplast, distribution of iron.) Z. Bot. 23. 957. Noack, K. and Liebich, H. 1941. (Chloroplast, distribution of iron.) Naturwiss. 29. 302. Noack, K. and Timm, E., 1942. (Chloroplast, amino acids.) Naturwiss. 30. 453. NoDA, H. and Wyckoff, R. W. G. 1951. (Rcprecipitated collagen.) Biochim. Biophys. Acta 7. 494. NowoTNY, H. and Zahn, H., 1942. (Keratin of wool, fine-structure.) Z. phys. Chem. B 51. 265. O'Brien, Jr. H. C, 1945. (Electron microscopy.) J. appl. Phys. 16. 370. Ogur,M.,Erickson,R.O., Rosen, G. U., Sax, K.B.andHoLDEN, C, 1951. (Ribonucleic acid in nucleus.) Exp. Cell Research 2. 73. Ohara, K., 1933a. (Silk, optics.) Sc. Pap. I.P.C.R. 21. 104. Ohara, K., 1933b. (Silk, swelling.) Sc. Pap. I.P.C.R. 22. 216. Ohara, K., 1938. (Wool, fine-structure.) Mclliand Textilber. 1938, No. 5. Ohara, K., 1939. (Wool, fine-structure.) Melliand Textilber. 1939, No. 5. Ohman, L., 1944. (Egg, lipo-protcic cortical layer.) Ark. f. Zool. (Stockholm) 36A. No. 7. Oltmanns, Fr., 1922. Morphologie und Biologie der Algen. Bd. i. 2. Aufl. Jena. OoRT, A. J. P., 1931. (Spiral growth.) Proc. Acad. Sci. Amsterdam 34. 564. OoRT, A. J. P. and Roelofsen, P. A., 1932. (Phycomyces, optics of sporangiophore.) Proc. Acad. Sci. Amsterdam 35. 898. OsTER, G., 195 1. (Long-range forces.) Z. Elektrochem. u. angew. physik. Chem. 55.529- OsTERGREEN, G., 1950. (Chromosomcs, displacement in mitosis.) Hereditas 36. i. OsTWALD, Wo., 1909. Grundriss der Kolloidchemie. Dresden. Ostwald, Wo., 1925. (Structural viscosity.) Kolloid-Z. 36. 99. OsTWALD, Wo., 1927. Die Welt der vernachlassigten Dimensionen. Leipzig. Ostwald, Wo., 1938. (Terminology in colloid science.) Kolloid-Z 84. 258. OvERBECK, F., 1934. (Cell elongation.) Z. Bot. 27. 129. Overton, E., 1899. (Lipid theory.) Viertelj. schr. Naturf. Ges. Zurich 44. 88. Pacsu, E., 1948. (Cellulose, structural problems.) Fortschr. Chem. org. Naturstoffe 5. 128. Painter, T. S., 1941. (Chromosomes, fibrillar structure.) Cold Spr. Harb. Symp. quant. Biol. 9. 47. 392 LITERATURE Pallmakn, H., 1931. (Hydration.) Viertelj. schr. Naturf. Ges. Zurich 76. 16. Pallmann, H., 1937. (Hydration.) Schweiz. med. Wochenschr. 67. 528. Palmer, K. J. and Hartzog, M. B., 1945. (Sodium pectate. X-ray diffraction.) J. Amer. Chem. Soc. 67. 2122. Patau, K., 1935. (Giant chromosomes.) Naturw'iss. 23. 537. Pauling, L., 1940. The nature of the chemical bond. 2nd ed. Ithaca and London. Pekarek, J., 1930. (Absolute viscosity in cells.) Protoplasma 11. 19. Pekarek, J., 1932. (Absolute viscosity in cells.) Protoplasma 17. i. Pekarek, J., 1933. (Phase separation in cell sap.) Protoplasma 20. 251. Perry, S. V., Reed, R., Astbury, W. T. and Spark, L. C, 1948. (Actomyosin, syneresis.) Biochim. Biophys. Acta 2. 674. Perutz, M. F., 1948. (Haemoglobin, molecular dimensions.) Nature 161. 204. Perutz, M. F., 1951. (Polypeptide chains, spiral configuration.) Nature 167. 1053. Perutz, M. F. and Mitchison, J. M., 1950. (Erythrocyte, sickle shape.) Nature 166. 677. Peterlin, a. and Stuart, H. A., 1943. (Induced double refraction.) Hand- u. Jahrb. d. Chem. Physik. Bd. 8. Abschn. ib. Leipzig. Peters, R. A., 1937. (Cell frame.) Perspectives in Biochemistn% p. 36 ff. Cambridge. Pfeffer, W., 1872. (Aleurone grains of Ricinus.) Jahrb. wiss. Bot. 8. 429. Pfeffer, W., 1877. Osmotische Untersuchungen. Leipzig. Pfeiffer, H., 1929. Elektrizitat und Eiweisse. Dresden. Pfeiffer, H., 1936. (Protoplasm, stuctural viscosit)-.) Physics U.S.A. 7. 302. Pfeiffer, H., 1937. (Cytoplasm, a non-Newtonian liquid.) Cytologia, Tokyo, Fujii Jubilee Vol. p. 701. Pfeiffer, H., 1941a. (Chromosomes, micro surgery.) Chromosoma 2. 77. Pfeiffer, H., 1941b. (Leptonics.) Ber. dtsch. bot. Ges. 59. 288. Pfeiffer, H., 1942a. (Leptones.) KoUoid-Z. 100. 254. Pfeiffer, H., 1942b. (Muscle, diffraction studies.) Protoplasma 36. 444. Pfeiffer, H. H., 1944. (Nuclear membrane, birefringence.) Z. wiss. Mikr. 59. 217. Philipp, H. J., Nelson, M. L. and Ziifle, H. M., 1947. (Cellulose fibres, crjstallinity.) Text. Res. J. 17. 585. Philippoff, \\"., 1935. (Structural viscosit\'.) Kolloid-Z. 71. i. Picken, L.E.R., 1949. (Chitin, proteinic precursor.) Phil. Trans. Roy. Soc. London B 234. I. Picken, L. E. R., Pryor, ]M. G. M. and Swann, M. IM., 1947. (Crossed lamellar systems.) Nature 159. 434. Pilnik, W., 1946. (Pectin, double refraction of flow.) Diss. E. T. H. Ziirich. Pirschle, K., 1930. (Series of ions.) Jahrb. wiss. Bot. 72. 335. Pischinger, a., 1937. (Nucleus, structure after fixation.) Z. Zellforsch. 26. 249. PiscHiNGER, A., 1950. (Nuclear structure.) Protoplasma 39. 567. Plowe, J. Q., 1931. (Plasmic membrane.) Protoplasma 12. 196, 221. Pochettino, O., 1913. (Hair, double refraction.) Atti Ac. Lincei. CI. Sc. Fis. (4) 22. 496, 696. PoLSON, A. and Wyckoff, R. W. G., 1947. (Haemocyanin, shape of molecules.) Nature 160. 153. Ponder, E., 1934. The mammalian red cell and the properties of haemolytic systems. Berlin 1934. Ponder, E., 1948. Haemolysis and related phenomena. New York. Posternak, S., 1929. (Nucleic acid as dissimilation product.) Bull. Soc. Chim. biol. Paris II. 787. Potter, V. R., 1944. (Myosin as adenosine triphosphatase.) Advances in Enzymol. 4. 201. LITERATURE 393 Pratt, A. W. and Wyckoff, R. W. C, 1950. (Collagen fibres, striation.) Biochim. Biophys. Acta 5. i66. Preston, J. M., 1933. Modern textile microscopy. London. Preston, R. D., 1934. (Tracheids, spiral texture.) Philos. Trans. B 224. 131. Preston, R. D., 1936. (Phycomyces, spiral growth.) Proc. Leeds Phil. Soc. 3. 327. Preston,' R. D., 1946. (Conifer tracheids, fine-structure.) Proc. Roy. Soc. London B 133- 327- Preston, R. D., 1947. (Crossed lamellar systems.) Proc. Roy. Soc. London B 134. 202. Preston', R. D., 1948. (Spiral growth in Phycomyces.) Biochim. Biophys. Acta 2. 155. Preston,' R. D., Nicolai, E., Reed, R. and Millard, A., 1948. (Valonia, microfibrils.) Nature 162. 665. Priestley, J. H., 1943. (Cuticle in Angiosperms.) Bot. Rev. 9. 595. Priestley, J. H. and Irving, A. A., 1907. (Chloroplasts, microscopic structure.) Ann. Bot. 21. 407. Prokofyeva Belgorskaja, a. a., 1948. (Genes, screened by heterochromatin.) J. Genetics 48. 80. Prudhomme van Reine, W". J., 1955. (Protoplasm, consistency.) Diss. Leiden. Rabinowitch, E. J., 1945. Photosynthesis and related processes, vol. L New York. Randall, J. T. and Friedlander, M.H.G., 1950. (Ram sperm, electron microscopy.) Exp. Cell Research I. I. Ranzi, S., 1951. (Viscosimetric distinction of fibrillar and globular proteins.) Experientia 7. 169. Raven, Ch., and Bretschneider, L. H., 1942. (Egg, stratification by centrifiigal force.) Arch. Neerl. Zool. 6. 255. Reinders, D. E., 1940. (Water resorption depending on respiration.) Diss. Groningen. Reinke, J., 1 88 1. (Plastin.) Unters. bot. Lab. Gottingen 2. i, 79. Renner, O., 1915. (Cohesion of water.) Jahrb. wiss. Bot. 56. 617. Reumuth, H., 1942. (Wool, fine-structure.) Klepzigs Textil-Z. 1942, p. 288. Rhumbler, L., 1898. (Cytoplasm as a liquid.) Arch. Entw. Mech. Org. 7. 103. Rhumbler, L., 1 914. (Protoplasm as a physical system.) Ergebn. Physiol 14. 474. Ries, E., 1938. Grundriss der Histophysiologie. Leipzig. Riley, D. P. and Oster, G., 195 i. (Nucleic acid, chain lattice.) Biochim. Biophys. Acta 7. 526. Roelofsen, p. a., 1949/50. (Phycomyces, spiral growth.) Rec. Trav. bot. neerl. 42. 72. Roelofsen, p. a., 1951a. (Phycomyces, wall texture of sporangiophore.) Biochim. Biophys. Acta 6. 357. Roelofsen, P. A., 1951b. (Cotton hair, primary wall.) Biochim. Biophys. Acta 7. 43. Roelofsen, P. A. and Kreger, D. R., 195 i. (Crystallized pectin in coUenchyma.) J. exp. Botany. 2. 332. RoMEis.B., 1943. Taschenbuch der mikroskopischen Technik. 14. Aufl. Miinchen u. Berlin Rosin, S., 1946. (Orthogonal fibrillar structure.) Rev. Suisse Zool. 53. 133. RoTHELi, A., Roth, H. and Medem, F., 1950. (Fish sperms.) Exp. Cell Research i. 115. RoTHEN, A., 1947. (Long-range forces extending over 200 A.) J. biol. Chem. 168. 75. Rothmund, V., 1898. (System phenol/water.) Z. phys. Chem. 26. 433. RoTTENBURG, W., 1 943. (Permeability to urea and glycerol.) Flora 37. 230. RozsA, G., Morgan, C, Szent-Gyorgyi, A. and W yckoff, R. W. G. 1950a. (Nerve, thin sections.) Science 122. 42. RozsA, G., Morgan, C, Szent-Gyorgyi, A. and Wyckoff, R. W. G., 1950b. (Myelinated nerve, electron mycroscopy.) Biochim. Biophys. Acta 6. 13. Rozsa, G., Szent-Gyorgyi,' a. and Wyckoff, R. W\ G., 1949. (Actin, linear agglo- meration.) Biochim. Biophys. Acta 3. 561, 594 LITERATURE Ro2sa,G.,Szent-Gyorgyi, A.and Wyckoff, R. W. G., 1950. (Myofibrils, fine-structure.) Exp. Cell Research 1. 194. RozsA, G. and Wyckoff, R. W. G., 1950. (Mitosis, electron microscopy.) Biochim. Biophys. Acta 6. 335. RozsA, G. and Wyckoff, R. W. G., 195 i. (Cytoplasm, electron microscopy.) Exp. Cell Research 2. 630. RucH, F., 1945. (Chromosomes, spiral structure.) Viertelj. schr. Natuf. Ges. Zurich 90. 214. RucH, F., 1949. (Meiotic chromosome, spiral structure.) Diss. E. T. H. Zurich; Chro- mosoma 3. 358. RuHLAND, W., 1912 (Ultrafilter theory'.) Jahrb. wiss. Bot. 51. 376. RuHLAND, W., 1950. (Permeability, ultrafilter theory.) Proc. VII th Int. Congr. of Botany, Stockholm. RuNDLE, R. E., Daasch, L. and French, D., 1944. (Starch, X-ray analysis.) J. Amer, Chem. Soc. 66. 130. RuNDLE, R. E., and Edwards, F. C. J., 1943. (Amylose, helical structure.) J. Amer. Chem. Soc. 65. 2200. RuNDLE, R. E., Foster J. F. and Baldwin, R. R., 1944. (Amylose, iodine reaction.) J. Amer. Chem. Soc. 66. 21 16. Runnstrom, J., 1928/29. (Egg, corticallayer.) Protoplasma 4. 388; 5. 201. Runnstrom, J., 1944. (Asterias egg, fertilization membrane.) Acta Zool. 25. 159. Runnstrom, J. and Monne, L., 1945. (Sea urchin egg, surface changes.) Ark. Zool. 36. A. No. 20. Runnstrom, J., Monne, L. and Wicklund, E., 1946. (Sea urchin egg, surface studies.) J. Colloid Sci. I. 421. Ruska, E., 1940. (Electron diffraction pattern in the electron microscope.) Wiss. Veroffentl. Siemens Werke Werkstoft'-Sonderh. Berlin 1940, p. 372. Ruska, H., 1939. (Electron microscope, preparation technics.) Naturwiss. 27. 287. Rutishauser, E., Huber, L., Kellenberger, E., Majno, G. and Rouiller, Ch., 1950. (Bone, electron microscopy.) Arch. Sci. (Geneve) 3. 175. Sadron, G., 1937. (Double refraction of flow, theory.) J. Phys. Radium 8. 481. Sakurada, I. and Hutino, O., 1933. (Silk fibroin, crystal lattice.) Sc. Pap. I.P.C.R. 21. 266. Samec, M., 1927. Kolloidchemie der Starke. Dresden. Samec, M., 1942. Die ncuere Entwicklung der Kolloidchemie der Starke. Dresden. Samec, M., 1942/43. (Starch, review.) Kolloidchem. Beih. 54. 147, 177, 435. Sauter, E., 1937. (Cellulose, X-ray diffraction.) Z. phys. Chem. B 35. 83, 117. Scarth, G. W., 1924. (Chloroplasts, double refraction.) Quart. J. exp. Physiol. 14. 99. ScARTH, G. W., 1927. (Plasm, microsurgery.) Protoplasma 2. 189. Scarth, G. W., 1942. (Cytoplasm, fine-structure.) In W. Seifriz, The structure of protoplasm. Ames-Iowa, p. 99. ScHAEDE, R., 1927. (Protoplasm, microscopic structure.) Protoplasma 3. 145. Schauenstein, E. and Stanke, D., 1951. (Collagen, UV absorption.) Makromol. Chemie 5. 262. ScHEiBE, G., 1948. (Protein crystals.) Naturwiss. 35. 168. Scherrer, a., 1914. (Chloroplasts and chondriosomes.) Diss. Univ. Zurich. Scherrer, p., 1920. (Particle size measured by X-rays.) Kolloidchemie von Zsigmondy, 3. Aufl. p. 387. Schimper, a. F. W., 1885. (Chloroplasts, grana structure.) Jahrb. wiss. Bot. 16. i. Schmidt, O., 1942. (Nerves, conduction of stimuli.) Naturwiss. 30. 644. Schmidt, W. J., 1923. (Tooth enamel, double refraction.) Ber. niederrhein. Ges. Natur- kunde Jahrg. 1923, Bonn 1925. LITERATURE 395 Schmidt, W. J., 1924. Die Bausteine des Tierkorpers im polarisierten Licht. Bonn. Schmidt, W. J., 1928. (Chromatin, double refraction.) Zool. Jalirb. 45. 177. Schmidt, W. J., 1934. (Submicroscopic structure of cells and tissues.) Handb. biol. Arbeitsmeth. v. Abderhalden, Abt. V. Teil 10, p. 435- Schmidt, W. J., 1936. (Ascaris, envelope of ovum.) Z. Zellforsch. 25. 181. Schmidt, W. J., 1936/37. (Tooth enamel, double refraction.) Naturwiss. 24. 361 ; 25. 283. Schmidt, W. J., 1937a. Die Doppclbrechung von Karyoplasma, Zytoplasma und Meta- plasma. Berlin. Schmidt, W. J., 1937b. (Nerve sheath, fine-structure.) Z. wiss. Mikr. 54. 159. Schmidt, W. J., 1937c. (Chromosomes, double refraction.) Naturwiss. 25. 506. Schmidt, W. J., 1938. (Form double refraction in sols.) Z. wiss. Mikr. 55. 476. Schmidt, W. J., 1939a. (Nuclear spindle, double refraction.) Chromosoma i. 253. Schmidt, W. J., 1939b. (Anisotropic chromosomes, microtechnics.) Z. wiss. Mikr. 56. i. Schmidt, W. J., 1939c. (Nuclear membrane, double refraction.) Protoplasma 32. 193. Schmidt, W. J., 1941a. (Chromosomes, double refraction.) Chromosoma 2. 86. Schmidt, W. J., 1941b. (Protplasm, double refraction.) Ergebn. Physiol, biol. Chem. u. Pharmakol. 44. 27. ScHMiTT, F. O., 1936. (Nerve fibre, ultrastructure.) Cold Spr. Harb. Symp. quant. Biol.4. 7. Schmitt, F. O., 1938. (Nuclear membrane, double refraction.) J. appl. Phys. 9. 109. ScHMiTT, F. O., 1950a. (Muscle, fine-structure.) Biochim. Biophys. Acta 4. 68. Schmitt, F. O., 1950b. (Nerve, myelin sheath.) Mult. Sclerosis & Desmyel. Diseases 28. 247. Schmitt, F. O. and Bear, R. S., 1937. (Nerve axons, optics.) J. cell. comp. Physiol 9. 261. Schmitt, F. O. and Bear, R. S., 1939. (Nerve axon sheath, fine-structure.) Biol. Rev. 14. 27. Schmitt, F. O., Bear, R. S. and Ponder, E., 1936. (Erythrocytes, optics.) J. cell. comp. Physiol. 9. 89. Schmitt, F. O., Bear, R. S. and Ponder, E., 1938. (Erythrocytes as composite bodies.) J. cell. comp. Physiol. 11. 309. Schmitt, F. O., Hall, C. E., and Jakus, M. A., 1942. (Collagen, electron microscopy.) J. cell. comp. Physiol. 20. 11. Schmitt, F. O., Hall, C. E., and Jakus, M. A., 1943. (Protoplasmic fibrils, fine-structure.) Biol. Symp. 10. 261. Schmitt, F. O., Hall, C. E., and Jakus, M. A., 1945. (Collagen fibrils, macroperiod.) J. appl. Phys. 16. 263. Schmitt, F. O. and PaImer, K. J., 1940. (Water absorption by myelin substances.) Cold Spr. Harb. Symp. quant. Biol. 8. 94. Schmitt, F. O. and Wade, L. J., 1935. (Nerve, thermal shortening.) Amer. J. Physiol. III. 159. Schmucker, Th., 1930. (Assimilation, quantum efficiency.) Jahrb. wiss. Bot. 73. 824. Schneider et al., 1936. (Pectin, esterification.) Ber. dtsch. chem. Ges. 69. 2530, 2537- Schoch-Bodmer, H., 1936. (Filaments, elongation.) Planta 25. 660. Schopfer, W. H., 1936. (Growth factors.) Arch. Mikrobiol. 7. 571. ScHOPFER, W. H., 1937. (Growth factors.) Ber. schweiz. bot. Ges. 47. 460. ScHULTZ, J., 1941. (Nucleic acids, importance in chromosomes.) Cold Spr. Harb. Symp. quant. Biol. 9. 55. ScHULZ, G. v., 1936. (Chain length of macromolecules, osmotic determination.) Z. phys. Chem. A 176. 317. ScHULZ, G. v., 1939. (Criticism of junction theory.) Z. angew. Chem. 52. 498. ScHiJRHOFF, P. N., 1924. Die Plastiden. Berlin. Schuringa,G. J.andAi,GERA,L., 1950. (Wool, epidermis.) Biochim. Biophys. .'\cta6. 325. 596 LITERATURE Scott, D. B. and Wyckoff, R. W. G., 1946/47. (Teeth, electron microscopy.) U.S.A. Public Health Rep. 61. 697, 1397; 62. 1513. Seifriz, W., 1927. (Erythrocytes, physical properties.) Protoplasma i. 345. Seifriz, W., 1929. (Cytoplasm, contractility.) Amer. Nat. 63. 410. Seifriz, W., 1936. Protoplasm. New York and London. Seifriz, W., 1937. (Protoplasmic flow.) Science 86. 397. Seifriz, W., 1938. (Protoplasm, fine-structure.) Science 88. 21. Seifriz, W., 1942. (Protoplasm, physical properties.) The structure of protoplasm. Mono- graph Amer. Soc. Plant Physiol. Ames-Iowa, p. 245. Seifriz, W., 1943. (Protoplasm, streaming.) Bot. Rev. 9. 49. Seifriz, W., 1945a. (Protoplasm, stnacture.) Bot. Rev. 11. 231. Seifriz, W., 1945b. (Torsion in protoplasm.) J. Colloid Sci. i. 27. Seifriz, W., 1950. (Anesthesis of protoplasm.) Anesthesiol. 11. 24. Seifriz, W. and Plowe, J.Q., 1931. (Cytoplasm, extensibility.) J. Rheology 2. Semmens and Bhaduri, 1939. (Nucleoli, differential staining with methyl green.) Stain Tech. 14. I. Senn, G., 1908. Gestaltveranderungcn der Pflanzenchromatophoren. Leipzig. Senti, F. R. and Witnauer, L. P., 1946. (A-starch, X-ray analysis.) J. Amer. Chem. Soc. 68. 2407. Serra, J. A. and Queiroz-Lopes, A., 1944. (Nucleolus, basic proteins.) Chromosoma 2. 576. Seybold, a., 1941. (Leaf pigments.) Bot. Arch. 42. 254. Seybold, a. and Egle, K., 1940. (Chlorophyll, physical state in chloroplast.) Bot. Arch, 41. 578. Seybold, A., Egle, K. and Hulsbruch, W., 1941. (Chlorophyll in algae.) Bot. Arch. 42. 239. Sharp, W. L., 1934. An introduction to cytology. 3rd ed. New York. Siedentopf, H. and Zsigmondy, K., 1903. (Ultramicroscope.) Ann. Phys. Lpz. (4) 10. i. Signer, R., 1930. (Double refraction of flow.) Z. phys. Chem. A. 150. 257. Signer, R., 1933. (Double refraction of flow.) Z. phys. Chem. A 165. 161. Signer, R., Caspersson, T. and Hammarsten, E., 193S. (Nucleic acids, double refraction of flow.) Nature 141. 122. Signer, R. and Gross, H., 1933. (Double refraction of flow.) Z. phys. Chem. A 165. 161. Sjostrand, F. S., 1949. (Retinal rods, layering.) J. cell. comp. Physiol. 33. 383. Smith, J. H. C, 1936. (Carotins, light absorption.) J. Amer. chem. Soc. 58. 247. Smith, J. H. C, 1937. (Leaf pigments.) Ann. Rev. Biochem. 6. 489. Snellman, O. and BjornstAhl, Y., 1941. (Birefringence of flow.) Kolloid. Beih. 52. 403. Snellman, O. and Erdos, T., 1948. (Myosin, molecular weight.) Biochim. Biophys. Acta 2. 650. Snellman, O. and Erdos, T., 1949. (F-actomyosin, actin-myosin quotient.) Biochim. Biophys. Acta 3. 523. Snellman, O. and Gelotte, B., 1950. (Muscle fibre, myosin extracted.) Exp. Cell Research i. 234. Soding, H., 1 93 1. (Cell wall, extensibility.) Jahrb. wiss. Bot. 74. 127. SoDiNG, H., 1934. (Cell wall structure of growing cells.) Jahrb. wiss. Bot. 79. 231. SoLLNER, K., 1950. (Selective permeability.) J. Electrochem. Soc. 97. 139. Speich, H., 1941. (Starch grains, optics.) Diss. E. T. H. Zurich. Spemann, H., 1936. Experimentelle Beitrage zu einer Theorie der Entwicklung. Berlin. Spoehr, H. a., 1935. (Leaf pigments.) Year Book Carnegie Inst. Wash. 34. 192. Sponsler, O. L., 1922. (Starch, X-ray diff'raction.) Amer. J. Bot. 9. 471. Sponsler, O. L., 1923. (Starch, crystal lattice.) J. gen. Physiol. 5. 757. LITERATURE 597 Sponsler, O. L. and Bath, J. D., 1942. (Protoplasm, molecular structure.) In W. Seifriz, The structure of protoplasm. Ames-Iowa, p. 41. Sprecher, a., Heierle, E. and Almasi, F., 1955- (Leaf pigments, quantitative spectrosco- py.) Biochem. Z. 283. 45. Stanley, W. M., 1935. (Crj'stallized virus protein.) Science 81. 644. Stanley, W. M., 1936. (Crystallized virus protein.) J. biol. Chem. 115. 673. Stanley, W. M., 1938a. (Virus protein.) Handb. d. Virusforsch. v. Doerr and Hallauer Bd. I, p. 447, Wien. Stanley, W. M., 1938b. (Progress in virus investigation.) The Sigma Xi Lextures. New Haven (Yale Univ. Press). Staudinger, H., 1932. Die hochpolymeren organischen Verbindungen. Berlin. Staudinger, H., 1936a. (Macromolecular chemistry.) Z. angew. Chem. 49. 801. Staudinger, H., 1936b. (Polymer homologous cellulose series.) ZellstofF-Faser 1956, Heft Nov./Dez. Staudinger, H., 1937a. (Cellulose, constitution.) Svensk kem. Tidskr. 49. 3. Staudinger, H., 1937b. (Cellulose, starch, glycogen.) Naturwiss. 25. 673. Staudinger, H. and Husemann, E., 1937. (Starch, constitution.) Lieb. Ann. 527. 195. Staudinger, H. and Sorkin, M., 1937a. (Cellulose, h^^drolysis.) Ber. dtsch. chem. Ges. 70. 1565. Staudinger, H. and Sorkin, M., 1937b. (Nitrocellulose.) Ber. dtsch. chem. Ges. 70. 1993. Staudinger, H., Staudinger, M. and Sauter, E., 1937. (Synthetic high polymers, microscopic structure.) Z. phys. Chem. B 37. 403. Steinbrinck, C, 1906. (Shrinking and cohesion.) Biol. Zbl. 26. 637. Steward, F. C, 1932. (Salt resorption and respiration.) Protoplasma 15. 497. Steward, F. C, 1933. (Salt resorption and respiration.) Year Book Carnegielnst. Wash. 32. 281. Stocker, O., 1928. (Halophytes.) Ergebn. Biol. 3. 265. Stoll, a., 1936. (Chlorophyll, constitution.) Naturwiss. 24. 53. Stoll, a. and Wiedemann, E., 1941. (Chloroplastin.) Verb. Schweiz. Naturf. Ges. Basel 1941, p. 125. Straub, F. B., 1948. (Actin, molecular weight.) Hung. Acta Physiol, i. 50. Straub, J., 1938. (Chromosomes, spiral structure.) Z. Bot. 33. 65. Straub, J., 1943. (Chromosomes, structure.) Naturwiss. 31. 97. Strugcjer, S., 1935/36. (Vital staining.) Protoplasma 24. 108, 26. 56. Strugger, S., 1936/37. (Chloropjasts, vital staining.) Flora 31. 113. 324. Strugger, S., 1949. Fluoreszenzmikroskopie und Mikrobiologie. Hannover. Strugger, S., 1950. (Proplastids.) Naturwiss. 37. 166. Strugger, S., 195 i. (Chloroplast, microscopic structure.) Ber. dtsch. bot. Ges. 64.69. Stuart, H. A., 1934. Molekiilstruktur. Berlin. Stuart, H. A., 1941. (Short-range order.) Kolloid-Z. 96. 149. Stubel, H., 1923. (Muscle fibre, rodlet double refraction.) Pfliig. Arch. 201. 629. Svedberg, Th., 1930. (Protein particles.) Kolloid-Z. 51. 10. Svedberg, Th., 1931. (Reserve protein, molecular weight.) Nature 127. 438. Svedberg, Th., 1938a. (Ultracentrifuge.) Industr. Engng Chem. 10. 113 (analyt. edition). Svedberg, Th., 1938b. (Proteins, multiple law.) Kolloid-Z. 85. 119. Szent-Gyorgyi, a., 1943. (Crystallized myosin). Studies Inst. Med. Chem. Univ. Szeged 3. 76. Takahashi, W. N. and Rawlins, T. E., 1933, 1935. (Virus protem, double refraction of flow.) Science 77. 26, 284; 81. 299. Takahashi, W. N. and Rawlins, T. E., 1948. (Tobacco mosaic virus, length.) Phytopath. 38. 279. 398 LITERATURE Tavel, p., 1939. (Cellulose, ester with dicarbonic acids.) Diss. Bern. Taylor, G. W., 1942. (Nerve, conduction.) J. cell. comp. Physiol. 20. 359. Teorell, T., 1935. (Selective permeability, theory.) Proc. Soc. exp. Biol. N.Y. 33. 282. Thimann, K. v., 1936. (Growth hormones, chemistry.) Curr. Sci. 4. 716. Thimann, K. V. and Bonner, J., 1933. (Cell walls in meristems, chemistry.) Proc. Roy. Soc. London B 113. 126. Thimann, K. V. and Bonner, J., 1938. (Growth hormones, chemistry.) Physiol. Rev. 18. 524. Thorell, B., 1948. (Erythrocytes, nucleic acids.) Cold Spring Harbour Symp. quant. Biol. 12. 247. TiMM, E., 1942. (Chloroplasts, amino acids.) Z. Bot. 38. i. Timofeeff-Ressovsky, N. W., 1940. (Mutation, biophysical analysis.) Nova Acta Leo- pold. 9. 209. Tischler, G., 1921/22. Allgemeine Pflanzenkaryologie. Berlin. ToRNAVA, S. R., 1939. (Plasmalemma, capacity of expansion.) Protoplasma 32. 329. Trillat, J. J., 1925/27. (Bimolecular films.) C.r. Acad. Sci. Paris 180. 1838, 184. 812. Trurnit, H. J., 1950. (Formvar film, penetrability of tr^'psin.) Science 112. 329. Tupper-Carey, R. M. and Priestley, J. H., 1923. (Cell walls of meristems.) Proc. Roy. Soc. London 95. 109. Ullrich, H., 1934. (Anion permeability.) Planta 23. 146. Ullrich, H., 1936a. (Protoplasm, double refraction.) Planta 26. 311. Ullrich, H., 1936b. (Cytoplasm boundary structure.) Protoplasma 26. 183. Ullrich, H., 1941. (Freezing of gels.) KoUoid-Z. 96. 348. Ursprung, a., 1915. (Cohesion of water.) Ber. dtsch. bot. Ges. 33. 153. Van de Sande Bakhuizen, H. L., 1925. (Starch grains, without layers.) Proc. Soc. Exp. Biol. N.Y. 23. 302. Van Iterson, G., 1927. (Cell wall, formation.) Chem. Weekbl. 24. 166. Van Iterson, G., 1933. (Cellulose symposium.) Chem. Weekbl. 30. i. Van Iterson, G., 1935. (Cell wall, fine-structure.) Proc. IV. Int. Bot. Congr. Amsterdam 2. 291. Van Iterson, G., 1937. (Protoplasmic flow and cell wall texture.) Protoplasma 27. 190. Van Iterson, G., Meyer, K. H., and Lotmar, W., 1936. (Plant chitin.) Rec. Trav. Chim. Pays-Bas 55. 61. Vermaas, D., 1941. (Adsorption double refraction.) Diss. Utrecht. Vermaas, D., 1942. (Adsorption double refraction.) Z. phys. Chem. B 52. 131. Verzar, F., 1943. Theorie der Muskelkontraktion. Rektoratsprogramm der Univ. Basel. Virtanen, a. I., 1948. (Cytoplasm, enzymes.) Oral communication. Vischer, E. and Chargaff, E., 1948. (Purines and pyrimidines, chromatography.) J. Biol. Chem. 176. 703. Vles, F., 1911. Proprietes optiques des muscles. Paris. Vogel, a., 1950. (Pollen tubes, electron microscopy.) Diplomarbeit am Pflanzenphysiol. Institut der E.T.H., unpublished. Vorlander, D., 1936. (Supracrystalline compounds.) Naturwiss. 24. 113. Wakkie, J. G., 1935. (Chlorophyll, light absorption.) Diss. Leiden. Walchli, O., 1945. (Cellulose, staining with congored.) Diss. E.T.H. Zurich. Warburg, O. and Negelein, E., 1923. (Assimilation, quantum efficiency.) Z. phys. Chem. 106. 191. Wassink, E. C, Vermeulen, D., Reman, G. H. and Katz, E., 1938. (Chlorophyll, fluorescence.) Enzymologia 5. 100. Waugh, D. F., 1950. (Erythrocyte, membrane.) Ann. New. York Acad. Sci. 50. 835. Weber, E., 1942. (Plant waxes, Qptics.) Diss. E.T.H. Zurich. LITERATURE 399 Weber, F., 1921. (Cell sap, viscosity.) Ber. dtsch. bot. Ges. 39. 188. Weber, F., 1932. (Plasmalemma, tonoplast.) Protoplasma 15. 453. Weber, F., 1933. (Chloroplast, myelin forms.) Protoplasma 19. 455. Weber, F., 1936. (Chloroplasts, grana.) Molisch-Festschr. (Wien), p. 447. Weber, F., 1937. (Anisotropic plastids.) Protoplasma 27. 280, 460. Weber, H. H., 1934. (Artificial myosin fibres, optics.) Pfliig. Arch. 235. 205. Weichsel, G., 1936. (Potato starch, resistance to enzymes.) Planta 26. 28. Weier, E., 1938. (Chloroplasts, review.) Bot. Rev. 4. 497. Weiss, P., 1939. Principles of development. New York. Went, F. W. and Thimann, K. V., 1937. Phytohormones. New York and London. Wergin, W., 1937. (Primary cell wall, optics.) Naturwiss. 25. 830. Wergin, W., 1942. (Cellulose fibres, electron micrograph.) KoUoid-Z. 98. 131. Werner, A., 1904. Lehrbuch der Stereochemie. Jena. Wettstein, F. von. 1937. (Cytoplasm as mediator of heredity.) Z. indukt. Abstamm. u. Vererb. lehre 73. 349. White, M. J. D., 1945. (Nucleus, chromosomes, genes.) In G. Bourne, Cytology and cell physiology. Oxford, p. 139. Wiegner, G., 1935. (Ion exchange.) Trans. 3. Int. Congr. Soil Sci. Oxford 3. 5. WiELER, A., 1936. (Chloroplasts, grana.) Protoplasma 26. 295. Wiener, O., 1912. (Theory of composite bodies.) Abh. sachs. Ges. Wiss. 33. 507. Wiener, O., 1926a. (Rodlet dichroism.) Ambronn-Festschr. Kolloidchem. Beih. 23. 189. Wiener, O., 1926b. (Incidental double refraction.) Ambronn-Festschr. Kolloidchem. Beih. 23. 198. WiLBRANDT, W., 1935. (Permeability.) J. gen. Physiol. i8. 933. Wilkins, M. H. F., Gosling, R. G. and Seeds, W. E., 1951. (Nucleic acid, optics.) Nature 167. 759. Williams, R. C. and Wyckoff, R. W. G., 1945. (Metallic shadow casting.) J. appl. Phys. 17. 31. Willstatter, R., 1933. (Theory of assimilation.) Naturwiss. 21. 252. ^X'illstatter, R. and Rohdewald, M., 1934. (Symplex theory.) Hoppe Seyler Z. physiol. Chem. 225. 112. WiNDAUs, A., 1923. (Sterins, molecular structure.) Hoppe Seyler Z. physiol. Chem. 130. 113. Winkler, K. C. and Bungenberg de Jong, H. G., 1940/41. (Erythrocytes, membrane.) Arch, neerl. Physiol. 25. 431, 467. Wirth, p., 1946. (Cell elongation.) Diss. E.T.H. Zurich. WissLER, A., 1940. (Birefringence of flow, technics.) Diss. Bern. Wohlfahrth-Bottermann, K. E., 1950. (Trichocysts of Paramecium.) Naturwiss. 37. 562. WoLPERS, C, 1941. (Erythrocytes, membrane.) Naturwiss. 29. 416. WoLPERS, C, 1944. (Collagen fibres, cross striation.) Virchows Arch. 312. 292. WoLPERS, C. and Ruska, E., 1939. (Blood fibrin, electron microscopy.) Klin. Wochenschr. 18. 1077, iiii. WoRSCHiTZ, F., 1935. (Muscle fibre. X-ray diffraction.) Fortschr. Rontgenstr. 51. 81. Wrinch, D. M., 1936. (Chromosomes, fine-structure.) Protoplasma 25. 550. Wrinch, D.M., 1937. (Cyclol theory.) Nature 139. 972. WuHRMANN, K., 1937. (Action of kations on cell elongation.) Diss. E.T.H. Zurich. WuHRMANN, K., Heuberger, A. and Muhlethaler, K., 1946. (Ultrasonics in electron microscopy.) Experientia 2. 105. WuHRMANN, K. and Pilnik, W., 1945. (Pectin, optics.) Experientia i. 330. Wuhrmann-Meyer, K. and M., 1939. (Avena coleoptile, cell wall structure.) Jahrb. wiss. Bot. 87. 642. 400 LITERATURE Wyckoff, R. W. G., 1947a. (Tobacco mosaic virus, electron micrographs.) Biochim. Biophys. Acta i. 143. Wyckoff, R. W. G., 1947b. (Globular virus, electron micrograph.) 6th. Int. Cong, of Cytology, Stockholm. Wyckoff, R. W. G., 1947c. (Long-range forces.) Biochim. Biophys. Acta 2. 139. Wyckoff, R. W. G., 1949. Electron Microscopy. New York-London. Wyckoff, R. W. G. and Corey, R. B., 1936. (Virus protein, X-ray analysis.) J. biol. Chem. 116. 51. Young, P., 1938. (Starch grains of Phaseolus and Canna.) Bull. Torrey bot. Club 65. i. Zacek, J. and Rosenberg, M., 1950. (Erythrocyte, electron microscopy.) Biochim. Biophys. Acta 5. 315. Zacharias, E., 1883. (Nuclein and plastin.) Bot. Ztg. 41. 209. Zernike, p., 1939. (Short-range order of molecules.) Z. Elektrochem. 45. 183. Zernike, p., 1946. (Phase contrast microscope.) in A. Bouwers Achievement in optics, Amsterdam, p. 116. Zetzsche, p., 1932. (Cutin.) Handb. d. Pflanzenanalyse 3. i. 205. Berlin. Ziegenspeck, H., 1938. (Turgor pressure mechanisms.) Bot. Arch. 39. 268. ZiMMER, K. G. and Timofeeff-Ressovsky, N. W., 1942. (Gene mutation by irradiation.) Z. indukt. Abstamm. u. Vererb. lehre 80. 353. ZiMMERMANN, A., 1 896. Morphologic und Physiologic des pflanzlichen Zellkerns. Jena. ZiMMERMANN, A., 1922. Die Cucurbitacecn. Jena. Zirkle, C, 1926. (Chloroplasts, microscopic structure.) Amer. J. Bot. 8. 301, 321. ZocHER, H., 1925. (Structure in sols.) Z. anorg. Chem. 147. 91. ZocHER, H., 1931. (Mesophases, optics.) Z. Kristallogr. 79. 122. Zollikofer, C, 1935. (Cell wall, extensibility.) Ber. dtsch. bot. Ges. 53. 152. ZsiGMONDY, R., 1921/22. (Definition of micelles.) Z. phys. Chem. 98. 14, loi. 293. ZsiGMONDY, R., 1925. Kolloidchemie. Leipzig. Zworykin, V. K., 1940. (Electron microscope.) Science 92. 51. ZwoRYKiN, V. K., 1 941. (Electron microscopy.) Cold Spr. Harb. Symp. quant. Biol. 9. 194. Zworykin, V. K., Hillier, J. and Vance, A. W., 1941. (Electron inicroscope.) Trans. electric, engin. 60. 157. Zworykin, V. K., Morton, G. A., Ramberg, E. G., Hillier, J. and Vance, A. W., 1945. Electron optics and the electron microscope. New York. AUTHOR INDEX Abbe, I, 119 Abe, 188 Adam, 49, 143 Algera, 256, 257, 338 Almasi, 250 Ambronn, 4, 76, 77, 85, 87, 89, 90, 96, loi, 293, 298, 308, 362 Anderson, 293, 307, 316 Andresen, 156 Ardenne, v., 116, 120, 123, 124, 127 Arisz, 199, 288 Arnold, 261 Arnon, 162 Aronoff, 251 Ascenzi, 350 Astbun,-, 158, 215, 229, 327, 328, 336, 339' 340, 342, 344, 346, 352, 353. 355, 356, 357, 364, 365, 366 Atkin, 366 Baas Becking, 252 Badenhuizen, 315, 316, 318 Bahr, 349, 369 Bailey, 328, 358, 369 Baker, 222 Baldwin, 312 Baltzer, 179 Bank, 24 Barlund, 199 Bartholome, 146 Bates, 312 Bath, 132 Baud, 217 Beams, 195, 222 Bear, 56, 86, 87, 270, 271, 319, 348, 361, 363, 364, 366, 367, 368 Becke, 315 Becker, 192, 218, 219, 230, 281 Belar, 222 Bell, 215, 229, 339, 340 Bensley, 172, 173, 184 Berger, 240 Bergman, B., 247, 261 Bergmann, AI., 134, 142, 335, 337 Bernal, 127, 138, 185, 241, 527 Bernfeld, 205, 311, 314, 315 Bernhard, 176 Bersin, 180, 208 Berthold, 184 Bessis, 172, 174, 176, 273 Bettelheim, 569 Beutner, 203 Beyer, 256 Bhaduri, 221 Biedermann, 306 Binz, 244 Biss, 176, 186 Bjomstahl, 92 Blackman, 261 Bladergroen, 156, 158 Blake, 364, 566, 367, 368 Blank, 287 Block, 343 Boehm, 90, 355, 361 Bogen, 205 Boissonnas, 313 Bolduan, 348, 368 Bonner, 22, 201, 209, 280, 281, 282, 288, 292 Borries, v., 116 Bosshard, 286 Bot, 246 Bottini, 80 Bougault, 296 Bouin, 174, 222 Bourdier, 296 Bourne, 173 Bragg, 97 Brandenberger, 350 Brauner, 149, 199, 297 Brenner, 173 Brcntano, 314 Bretschneider, 122, 174, 175, 274, 275, 277 Bricka, 174, 176, 273 Brill, 337 Brown, D.E.S., 164, 172 Brown, G. L., 217 Brown, H. P., 276 Broyer, 199 Bull, 51 Bungenberg de Jong, 18, 19, 20, 22, 23, 24, 265, 266, 267, 268, 269 Burstrom, 199, 288 Burton, 115 Biissem, 81 Biitschli, 182 Callan, 217, 218 Caspersson, 173, 212, 213, 214, 219, 220, 221, 227, 229, 238, 239, 353, 357 Castle, 289, 290, 304, 305, 308 Chambers, 163, 199 Champetier, 346, 366 Champy, 174 Chargaff, 214 Chibnall, 246, 292, 295, 296 Chodat, 170 Cholodny, 1 5 1 Christiansen, 286 Clark, 327 Clarke, 276 Claude, 122, 172, 173, 174 Cohn, 134 Collander, 197, 198, 199 Comar, 246 Corey, 241 Correns, 290 Corrigan, 327 Crick, 171 Crowfoot, 327, 328 Curtis, 199 Czaja, 100 Daasch, 319 Danielli, 197, 199, 200 Danon, 122 Davson, 197 Debye-Scherrer, 99, 100, loi, 104, 106, 319, 327 De Leeuw, 105 402 AUTHOR INDEX Denny, 210 Derkscn, 318, 344 De Rooy, 366 Dervichian, 263, 269 Dettmer, 349 Deuel, 60, 64, 68, 73, 201 Devaux, 5 1 Dickinson, 328, 355, 356 Diehl, 301, 309 Dobry, 21 Donnan, 75, 202, 203, 204 Doutrelignc, 243, 244 Drawert, 161, 193, 205 Drucker, 335 Dumpert, 312 Ebner, v., 347 Eckerson, 317 Edsall, 134, 354 Edwards, 312, 320 Egle, 251, 252 Einstein, 10 Elod, 342 Elvers, 231 Emerson, 261 Engel, 280 Engstrom, 226, 363 Erdos, 352, 358 Erickson, 215 Ernst, A., 256 Ernst, R., 217 Euler, v., 247, 261 Evans, 222 Eymers, 261 Fajans, 30 Fankuchen, 127, 185, 327 Farr, 317 Faure-Fremiet, 172, 173, 174, 346, 359, 366 Fehling, 310 Felix, 211 Fernandez-Moran, 362, 363, 364 Feulgen, 176, 214, 215, 221, 224, 226, 227 Fischer, A., 275 Fischer, E. H., 314 Fischer, F. G., 213, 214 Fischer, H., 248 Fitting, 210 Flaig, 363 Flcmming, 182, 192, 255 Foster, 93, 276, 312 Fournet, 263, 269 Franck, 261 Frank, 35 1 French, 312, 319 Freudenberg, 33, 312- Freundlich, 9, 18, 42, 66, .172 Frey, R., 128, 301, 308 Fricke, 268 Friedel, 52 Friedlfinder, 275 Friedrich-Freksa, 337, 366, 368 Fritz, 293 FuUam, 122, 173, 174 Gaffron, 261 Gaumann, 301 Gause, 207 Gautier, 176 Geitler, 217, 225, 240, 243 Gelotte, 354 Gerendas, 354, 358 Gerngross, 69, 345 Gettner, 122 Gibbons, 313 Gibbs-Thomson, 44, 48 Gicklhorn, 55, 221 Giroud, 173 Goebel, i Goldacre, 186 Goldschmidt, 29, 148, 149 Gorter, 51, 143, 268, 270 Gosling, 228 Gough, 263, 264 Graham, 8, 326 Gralen, 338 Granick, 245, 256, 257, 258, 265 Grendel, 268, 270 Grignolo, 176 Gross, H., 9 Gross, J., 176, 348, 351 Guilliermond, 8, 173, 191, 243, 244, 245 Guinier, 263, 269 Gundermann, 280, 292 Haas, 354 Haase-Bessel, 236 Hadorn, 180 Hagen-Poiseuille, 63, 65, 66 Hakansson, 222 Hall, 1 18, 348, 349, 356, 357,358,368 Halle, 34, 346 Hammarsten, 214, 220, 229 Hanes, 312, 314 Hanson, 316, 317 Hansteen-Cranner, 292 Harder, 301 Hargitay, 359 Harris, 216 Hartzog, 61 Harvey, E. B., 195, 223, 243 Harvey, E. N., 165, 166, 194, 200 Haurowitz, 331 Hausermann, 295 Haworth, 38 Hecht, 170 Hegetschweiler, 334 Heidenreich, 120 Heierle, 250, 251 Heilbronn, 167, 168, 169 Heilbrunn, 146, 147, 163, Heinrich, 314 Heitz, 217, 219, 222, 225, 226, 239, 243, 244, 245, 257, 258, 261 Hellstrom, 247, 261 Helly, 174 Hengstenberg, 34, 98 Heringa, 344 Hermans, 105, 106, 107, 108, 109, no, 112, 113, 114 Herrmann, 69 Hertwig, 176 Herzog, 78 Hess, 121, 280, 292 Heuberger, 121 Hevn, 280, 287 Hill 3 59 Hillier, 116, 119, 122 Hirschler, 210, 240 Hoagland, 199 Hober, 41, 43, 148, 202 Hoerr, 173 Hofler, 184, 197, 198, 200, 205 Hofmeister, 148, 151, 153, 199, 243 Hogeboom, 173 Hohnel, v., 338, Holden, 2 1 5 Holmstrom, 198 Hotchkiss, 173 Houwink, 276 Huber, 350 Hubert, 243, 252 Hughes, 171 Hiilsbruch, 251 lliirthle, 353 Husemann, 11, 118, 311, 317 AUTHOR INDEX 40 ; Huskins, 226 Husted, 124 Hutino, 337 Huxley, 359 Induni, 116, 117 Irvine, 313 Irving, 245 Itschner, 302 Jaag, 301 Jaccard, 95 Jacob, 247 Jaeger, 227 Jakus, 118, 348, 349, 356, 357, 358, 368 Jaloveczky, 323 Jeener, 173 Jensen, 275 Joly, 51, 93> 176 Jordan, 329, 331 Jorpes, 265 Jung, 264, 269, 273 Kamiya, 187, 188 Karrer, 208, 249, 250, 265, 300 Karstens, 256 Katz, E., 251 Katz, J. R., Ill, 316, 317, 318, 345, 366 Kausche, 256, 257 Kekule, 58 Kelansy, 317 Kellenberger, 122, 350 Kellner, 219 Kerr, 307, 316 Kiesel, 140, 183, 210, 211, 214., 238 Kiessig, 280 Kinsinger, 119 Knapp, 215 Knoch, 369 Koepfli, 209 Kohl, 1 1 5 Kohler, i Konig, 369 Koning, 252 Kopp, 31 Kopscli-Regaud, 174 Kossel, 211 Koydl, 256 Kozbial, 230 Kratky, 78, 89, 106, 107, 108, 115, 312, 337, 366, 368 Kreger, 280, 295, 319, 320 Kuhn, R., 250 Kuhn, W., 359 Kundt, 253, 298 Kiintzel, 72, 347 Kuriyama, 337, 366 Kiister, 171, 194, 220, 221, 243, 244, 253, 256 Kuwada, 230 Lagermalm, 344 Lamm, 314 Landolt-Bornstein, 169 Langelaan, 353 Langmuir, 49 Laue, V. 98 Lea, 232 Leaf, 217 Lecher, 42 Lehmann, F. E., 172, 174, 176, 179, 186, 190, 195, 196, 200, 276, 277, 278 Lehmann, O., 17 Lepeschkin, 264 Leuthardt, 173 Leuthold, 327 Levan, 222 Lewis, C. M., 261 Lewis, E. B., 239 Lewis, W. H., 186, 187 Liebich, 247 Liechti, 232 Lieser, 63, 68 Lindau, 42 Lindberg, 338, 344 Lindley, 343 Lindstrom, 363 Loeb, 203 Loewy, 189 Lomax, 327 Loomis, 194 Lorand, 369 Lorch, 186 Loschmidt, 27, 142, 167 Lotmar, 301, 302 Liidtke, 334 LundegS,rdh, 182, 199 Liischer, 295, 296, 297 Liithy, 363 Luyet, 196, 217 McArthur, 369 McClintoc, 239 Mackinney, 250 Magnus, 29 Majno, 350 Manegold, 76 Mangenot, 8, 173, 243 Manton, 276 Mark, 30, 31, 32, 38, 50, 59, 77, 78, 97, 98, 302, 310, 314, 329, 335, 337, 341, 348, 359, 367 Marklund, 199 Marks, 364, 366, 367, 368 Alarsland, 164, 172, :86 Martens, 215, 300 Martin, 116 Marwick, 344, 366 Matheson, 120 Matoltsy, 176, 354, 358 Matthey, 226 Maxwell, 92 Mazia, 227 Medem, 275 Meeuse, 281 Menke, 140, 245, 246, 247. 248, 249, 253, 255, 256, 257, 258, 260 Mercer, 337 Mestre, 247 Aletz, 233 Metzner, 245 Meyer, A., 243, 326 Meyer, K.H., 30, 31,32, 38, 50, 59, 68, 77, 78, 97, 98, 142, 185, 201, 203, 204, 205, 301, 302, 310, 311, 314, 315, 335, 337, 348, 359, 360 Meyer, K. P., 252 Meyer, M., 293, 294, 299, 300 Michaelis, 156 Michel, 124 Michener, 210 Middlebrook, 343, 369 Millard, 307 Mikier, 250 Minder, 232 Mirsky, 184, 191 Misch, 98 Mitchison, 200, 264, 270, 271 Mittasch, 40 Mohring, 77, 308 Mollendorf, v., i8i Mommaerts, 245, 247 Monne, 170, 173, 176, i8i, 182, 190, 217, 277 Monroy, 200, 276 Montalenti, 276 Moore, 316 Morgan, C, 364 404 AUTHOR INDEX Morgan, T. H., 230 Morton, 100 Mothes, 1 5 5 Moyer, 186 Miihlethaler, 121, 127, 128, 129, 173, 256, 259, 282, 283, 284, 304, 308 Muller, II, 233, 234, 255, 335 Muller, A., 34 Muller, A. F., 173 Muller, H. O., 120, 122, 124, 125 Muralt, v., 353, 354, 362 Myrback, 313, 314 Nageli, 4, 76, 77, 78, 79, 80, 108, 315, 326, 327 Nageotte, 55, 57 Nakamura, 230 Naray-Szabo, 319 Nathanson, 197 Nebel, 225 Neckel, 349 Needham, 156, 157, 359 Negelein, 261 Nelson, 113 N6mec, 216 Newman, 122 Newton, 164, 170 Nicolai, 252, 307 Niemann, 134, 142, 335, 337 Niggli, 27 Noack, 245, 247, 252 Noda, 349 Nowotny, 342 Oberling, 176 O'Brien, 121 Ogur, 215 Ohara, 331, 332, 333, 334, .. 339 Ohman, 276 Oltmanns, 169 Oort, 303, 304, 305 Oster, 177, 215 Ostergreen, 224 Ostwald, 8, 16, 18, 66 Overbeck, 281 Overton, 197, 198 Pacsu, 68 Painter, 230 PaUmann, 31, 149 Palmer, 57, 61, 361 Pankow, 302 Pasewaldt, 124, 125 Pasteur, 207 Patau, 227 Patterson-Fourier, 328 Pauling, 147 Pekarek, 167, 168, 169, 191 Perr}% 352 Perutz, 135, 265, 270, 359 Peterlin, 91, 108 Peters, 181 Peyer, 300 Pfeffer, 76, 327 Pfeiffer, 79, 154, 167, 168, 170, 195, 217, 229, 353 Philip, 338, 344 Philipp, 113, 329, 341, 367 Philippoff, 66 Picken, 302, 307, 310, 360 Piguet, 314 Pilnik, 61, 92 Piper, 292, 295, 296 Pirschle, 153 Pischinger, 193, 212, 215, 216, 218 Plantefol, 8, 173, 243 Platzek, 115 Ploetz, 312 Plowe, 170, 205 Pochettino, 338 Poiseuille, 170 Pollister, 184 Poison, 125, 126 Ponder, 167, 169, 262, 263, 264, 270, 271, 273 Porter, 256, 257 Posternak, 238 Potter, 359 Prakke, 347 Pratt, 348, 349 Preston, J. M., 333 Preston, R. D., 95, 305, 307 Priestley, 245, 280, 300 Prokofyewa Belgorskaja, 239 Prudhomme van Reine, 170 Pryor, 307 Queiroz-Lopes, 221 Rabinowitch, 246, 261 Randall, 218, 275 Ranvier, 364 Ranzi, 136, 137 Raven, 277 Rawlins, 127, 241 Reed, 307, 353 Reinders, 199 Reinke, 183, 184 Reman, 251 Renner, 42 Reumuth, 338 Rhumbler, 163, 164, 170 Ries, 156 Riley, 215 Ringer, 56, 166 Roelofsen, 280, 285, 303, 304, 305 Rohdewald, 247 Romeis, 221 Rosen, 215 Rosenberg, 273 Rosin, 307, 308 Rossenbeck, 214 Roth, 275 Rotheli, 275 Rothen, 159 Rothmund, 46 Rottenburg, 205 Rouiller, 305 Rozsa, 174, 2i6, 218, 222, 352, 357, 358, 364 Ruch, 226 Ruhland, 197 Rundle, 312, 319, 320 Runnstrom, 170, 190, 276 Ruska, E., 72, 116, 120, 129, 368 Ruska, H., 10, 118, 121, 256, 257, 349 Rutishauser, 350 Ruttle, 225 Sadron, 92, 1 1 5 Sakurada, 337 Samec, 314, 315 Samsa, 93 Sankewitsch, 1 5 1 Sauter, 98, 318 Sax, 215 Scarth, 168, i8i, 186, 199, 200, 253 Schaaf, 312 Schaede, 216 Schardinger, 312 Schauenstein, 337, 345, 346 Scheibe, 135, 329, 330 Scherrer, A., 256 Scherrer, P., 17, 77, 98 Schiff, 214 Schimper, 243 Schinz, 350 Schmidt, O., 362 Schmidt, W. J., 76, 92, 96, 178, 217, 220, 222, 224, AUTHOR INDEX 405 228, 229, 230, 255, 274, 303, 304, 306, 317, 347, 351. 353. 354. 355. 361, 362 Schmiedeberg, 210 Schmitt, F. O., 56, 57, 86, 87, 118, 217, 229, 270, 271. 348, 349. 356, 357, 358, 361, 362, 363, 364, 368 Schmucker, 261 Schneider, 64 Schneidmcsser, 312 Schoch-Bodmer, 281 Schopfer, 243 Schossberger, 31 Schultz, 240 Schulz, G. v., 61 Schulz, J., 185 SchiirhofF, 243 Schuringa, 338 Schweizer, 62, 65, 300 Schwendener, 76, 77 Scott, 351 Searle, 317 Seeds, 228 Seifriz, 67, 77, 170, 171, 172, 182, 186, 187, 263, 264 Sekora, 337, 368 Semmcns, 221 Senarmont, 294 Senn, 243, 257 Senti, 319 Serra, 221 Seybold, 251, 252 Sharp, 243 Siedentopf, 4, 8 Sievers, 203 Signer, 9, 90, 214, 220, 229, 298 Sjostrand, 363 Smith, J. H. C, 250 Smith, S. G., 355 Snellman, 92, 352, 354, 35 8 Soding, 281, 287 Sollner, 201 Sorkin, 65, 317 Spark, 353 Speich, loi, no, 113, 320, 321. 324 Spemann, 178 Spoehr, 250 Sponsler, 132, 310, 319 Sprecher, 250 Stanke, 345, 346 Stanley, 11, 127, 240 Stapelfeldt, 9 Staudinger, 5, 58, 59, 61, 62, 63, 64, 65, 67, 68, 78, 157, 311, 317. 318 Stecher, 286 Steinbrinck, no Steinmann, 254, 257 Steward, 199 Stocker, 153 Stokes, 10, 167, 168, 170, 191 Stoll, 247, 248, 260, 261 Stout, 162 Straub, F. B., 352 Straub, J. 220, 226 Strugger, 162, 193, 245, 247, 256, 257, 258 Stuart, 31, 91, 108 Stubel, 77, 355 Svedberg, 10, 11, 61, 125, 141, 142, 144, 159, 231, 260, 265, 331, 352 Swann, 307 Szent-Gyorgyi, 352, 357, 358, 364 Takahashi, 127, 241 Tavel, 68, 73 Teorell, 201, 266 Thaureaux, 172 Thimann, 209, 280, 286, 292 Thorell, 353, 357 Thung, 256 Timm, 247 Timofeeff-Ressovsky, 231, ,232, 233, 234, 235 Tischler, 219, 221 Tomlin, 218 Tornava, 199 Treer, 107 Trillat, 56 Trogus, 280 Trurnit, 159 Tupper-Carey, 280 Tyndall, 8 Ullrich, 90, 178, 199, 202, 253 Ursprung, 42 Vance, 1 1 6 Van de Sande Bakhuizen, 316 Van der Waals, 31, 35, 61, 93, 159, 185, 266 Van Dijk, 199 Van Iterson, G., 281, 290, 291, 301, 307, 309 Van Iterson, W., 256 Van 't Hoff, 61 Vermaas, 112, 115 Vermeulen, 251 Verne van Bremen, 222 Verzar, 559 Virtanen, 179 Vischer, 214 Vl^s, 353 Vorlander, 52, 53 Wade, 364 Wiilchli, 103, 304 Wakkie, 252, 253 Warburg, 261 Wassink, 251, 261 Waugh, 50, 264 Weber, E., 90, 299, 300 Weber, F., 191, 192, 205, 245, 247, 253 Weber, H. H., 355, ^(64 Weibull, 365, 366 Weichsel, 324 Weidinger, 113, 344 Weier, 243 Weiss, 179 Went, 209 Wergin, 121, 280, 281, 292 Werner, 28, 33 Wettstein, v., 242 Weurman, 252 Weyl, 81 White, 220 Wicklund, 170 Wiedemann, 248 Wiegner, 80 Wieler, 244, 245 Wiener, 82, 83, 84, 88, 89, loi, 115, 220, 270, 334, 338, 350. 351. 355 Wilbrandt, 202 Wilkins, 228 Williams, 125, 126 Willstiitter, 247, 261 Windaus, 138 Winkler, 265, 266, 267, 268, 269 Wirth, 287 Wissler, 92, 220 Witnauer, 319 Wohl, 261 Wohlfahrt-Bottermann, 368 Wolpers, 72, 129, 264, 268, 271, 272, 348, 357, 368 Worschitz, 355 Wrinch, 158, 227, 329 Wuhrmann ,61, 121, 287 4o6 AUTHOR INDEX Wuhfmann-Meyer, 292 WyckofF, 115, 121, 122, 123, 125, 126, 127, 129, 159, 174, i75> 216, 218, 222, 241, 256, 257, 282, 308, 348, 349, 351, 352, 357, 358, 364 Young, 316, 363, 364 Zacek, 273 Zacharias, 184 Zahn, 342, 343 Zegar, 300 Zenker, 255 Zernike, i, 108 Zetzsche, 296, 297 Ziegenspeck, no Ziifle, 113 Zimmer, 231 Zimmermann, 221, Zirkle, 245 Zocher, 9, 52, 53 ZoUikofer, 287 Zollinger, 173 Zsigmondy, 4, 8, 17, 76 Zworykin, 115, 116 327 10, II, I SUBJECT INDEX absorption, anisotropic, 85 light, 252 ultraviolet, 219, 226 achromatin, 219 actin, 352 active elimination, 198 active group, 208 active plasma, 181 actomyosin, 352 adenoid activity, 198 adenosine triphosphate, 358 aerogel, iii aggregation, 159, 330 agon, 236 aleurone grains, 193, 326 aliphatic compounds, 34 amino acids, 132, 330 amylase, 314 amylopectin, 310 amy lose, 310 angle of scattering, 90 anisotropic absorption, 83 anisotropic diffraction, 83 anisotropy, optical, 85 anucleal, 224 apo-enzyme, 208, 235 apposition growth, 290, 316 arginine, 340 aromatic compounds, 36 assimilation, unit of, 261 atomic distances, 29, 30. autocatalytic reproduction, 241 auto-reproduction, 240 auxins, 209 axolemma, 364 axon, 363 backbone spacing, 336, 365 bast fibres, 105 beaded chains, 93, 176 Bergmann-Niemann rule, biochemistry, 7 biomorphology, 7, 372 biosomes, 174 birefringence (cf. double re- fraction), 83, 85, 114 form, 85, 254 intrinsic, 88, 298, 308 layer, 255 lamellar = layer of flow, 90, 298 platelet = layer rodlet, 84, 114, 309, 321 block structure, 316 blood corpuscles, red s. ery- throcytes bones, 350 boundary layers, 197 cap-plasmolysis, 184, 197 carboxylase, 236 carotenes, 249 carotenoids, 246 carrier hypothesis, 233 cell elongation, 281 cell extension, 282 cell polarity, 190 cell wall, cutinized, 293 meristematic, 279 micellar textures, 95 primary, 279 secondary', 279, 290 cellobiose, 39 cellodextrines, 62 cellulose, 59, 62, 97, 280 cellulose frame, 286 centrifuge microscope, 194 centromeres, 222, 224 chain lattice, 34, 36, 69 chain length, 61 chains, beaded, 93, 176 chitin, 130, 301, 308 chlorophyll, 246, 248, 252 chloroplastin, 248, 258 chloroplastin symplex, 247 chloroplasts, 243 cholesterol, 138, 266 choline, 138 chondrioconts, 173 chondriome, 173 chondriosomes, 173 chromatid threads, 225 chromatin, 219 chromidia, 176 chromocentres, 220 chromomeres, 225 chromonema theor)', 225 chromonemata, 225, 241 cbromophily, 219 chromoprotein, 248, 249, 258, 265 chromosomes, 217, 224, 231, 234 coacervation, 18 coagulation, directed, 183 co-enzyme, 208, 235 cohesion, 31 molar, 32 cohesive bonds, 146, 185 cohesive forces, 31, 32, 185 Van der Waals, 31 cohesive pressure, 42 cohesive tension, 42 coleoptile, 287 collagen, 307, 345 collagen fibres, 348 collagen group, 367 colloid chemistry, 4 colloid particles, 8, 15 colloid solutions, 64 colloids, 15 corpuscular, 74, 81 globular, 126 reticular, 74, 81 composite bodies, 82 condensation, 59 connective tissue, 345 constellation, 37 contractility, 134 contraction, muscular, 358 of protoplasm, 187 co-ordination, 28, 29 corpuscular colloids, 73, 8 1 cotton fibres, 307 crossed lamellar systems, 506 crystal lattice, 26, 69, 97 crystal structure, 26 4o8 SUBJECT INDEX crystalline liquid, 52 crystalline, super-, 53 crystalloids, 194, 221, 326 crystals, liquid, 51 cuticle, 293 cuticular layers, 293, 297 cuticular transpiration, 301 cutin, 293, 297 cutin waxes, 294 cutinized cell walls, 293 cyanophily, 219 cyclic compounds, 37 cysteine, 133, 155 cystine, 155, 340 cytoplasm, 132 behaviour of proteins, 141 different phases, 191 fixation, 174 flow, 186 ground-, 174, 179 heredity, 242 molecular constituents, 132 molecular morphology, 207 morphogenesis, 179 permeability, 197 physical properties, 163 submicroscopic structure, 172 surface tension, 165 viscosity, 146, 166 water content, 177 cytoplasmic layers, 200 cytoplasmic membrane, 199 Debye-Scherrer diagram, 99 deflection discs, 9 dehydration, 20, 177 dehydrogenase, 207, 236 dehydrogenation, 154 denaturation, 136, 143, 329, 330 deplasmolysis, 198 dermatosomes, 317 desoxyribose, 212 desoxyribose nucleic acids, 237 dextrins, 312 dialysis, 8 diamino acids, 133 dichroism, 83, loi difi^iaction, anisotropic, 83 dipeptide, 132 dipole, 19, 147 disaccharides, 39 dispersed phase, 15 dispersing medium, 1 5 dispersion series, 70 dispersoids, 15, 16 Donnan equilibrium, 202 double refraction (cf. bire- fringence), 83, 84, 114 incidental, 88 intrinsic, 88, 298, 308 orientation, 89 tension, 89 ectoplasm, 186, 199 eggs, 190, 194, 276 elastic tissue, 351 elasticity, 65, 171 of flow, 164 structural, 164 elastin, 351 elastoidin, 346 electron diff^raction dia- grams, 120 electron microscope, 3, 116 electron microscopy, 115, 123 cell wall, 283 chloroplasts, 256 erythrocytes, 271 gels, 127 globular colloids, 1 26 muscle fibres, 357 electron rays, 115 electrophoresis, 8 electrosmosis, 75 elementary cell = unit c, 26, 97, 319 elimination, active, 198 elongation growth, 288 enamel, 350 enchylema, 181, 215, 217 endo-enzyme, 208 endoplasm, 186, 199 energy equivalent, 30 enzymes, 180, 207, 235 epidermis, 293 equilibrium liquid, 21 erythrocytes, 262 erythrophily, 219 estolids, 296 euchromatin, 220 extension growth, 283 extinction angle, 91 fat, 137, 139 fatty acids, 137 feather keratin, 344 Feulgen's nucleal reaction, 214 fibre diagram, 97, 99, 104 fibre texture, = fibrous t., 94, 290, 303, 332 fibres, 1 01 bast, 105 cellulose, 105 collagen, 348 cotton, 307 muscle, 352 ramie, 97, loi, 104 fibrillar hypothesis, 227, 237 fibrillar proteins, 135, 364 fibrillar theory, 182 fibrils, origin of, 194 protoplasmic, 194 fibrinogen, 369 fibroid texture, 94, 304, 332 fibroin, 331 fibroinogen, 337 fibrous texture = fibre t. fine-structure, 5 finger nails, 344 fixation, 75, 161, 174 flagella, hairy, 275 Flimmergeisseln, 275 flow, birefringence of, 90, 298 protoplasmic, 164, 186 flow-birefringence appar- atus, 299 fluorescence, 162, 252 focal depth, 120 foliate texture, 96, 304 form birefringence, 85, 254 framework, 176 gel, 66, 68, 69 mi cellar, 73, 81 molecular, 73, 80, 81, 201, 341 free rotation, 37 freeze -drying, 178 freezing, 196 galacturonic acid, 60 gametes, 274 gel framework, 66, 68, 69 gel-sol transition, 171, 187 gel solutions, 65 gels,_5 8, 71, 82 cristallinity, 112 electron microscopy, 127 polarization microscopy, 82 structure, 58, 66 swelling, 109 X-ray analysis, 96 gelatin, 21, 23, 93, 345 1 i SUBJECT INDEX 409 genes, 230, 233, 240 ghost, 264 Gibbs-Thomson theorem, 44 glasses, 81, 112 globoid, 194 globular molecules, 135 globular proteins, 135, 141 globulins, 212 glucosamine, 301 glucose, 37, 60, 313 glucosidases, 40 glucosides, 40, 310 glutamic acid, 340 glutathione, 155 glycine, 336, 346 glycogen, 3 i i grana, 245, 247, 257 growth, apposition, 290, 316 elongation, 288 extension, 283 forces of, 287 in area, 282, 307 mosaic, 286 spiral, 304 substances, 209 surface, 282, 307 tip, 282 guanyl nucleic add, 214 haemocyanin, 125 haemoglobin, 265, 269 haemolysis, 264 Haftpunkt-Theorie, 145 Haftpunkte, 67 Hagen-Poiseuille's law, 65 hair, 338 hairy flagella, 275 Hecht, strands of, 170 hemicelluloses, 61 heredity, 230, 242 heterocapillarity, 105 heterochromatic parts, 220 225, 239 heterocyclic, 37 heterogeneous, 12 heteropolar lattice, 28 hexane, 36 high polymers, 58 histidine, 210 histones, 212 homocapillarity, 105 homogeneity, optical, 12 physico-chemical, 12 homogeneous, statistically, 12 homopolar lattice, 28 honeycomb theory, 182 hormones, 208 horny substances, 338 hydration, 19, 148, 163 hydrogen, 208 hydrogen bonds, 147 hydrogen bridges, 147 hydrogen pressure, 156 hydrogenation, 155 hydrolysis patterns, 317 hydrophilic groups, 47 hydrophily, 44 hydrophobic groups, 47 identity period, 26 I.E. P. = isoelectric point, 153. 154 _ imbibition liquid, 84 imbibition water, 196 incidental double refraction, 88 index ellipsoid, 87 insulin, 328 interferences, 97 intermicellar, 79, 81 intermicellar phase, 79 intermicellar portion, 79 intermicellar processes, 80, 81 intermicellar spaces, 99, 100 intermicellar substances, 81 intermicellar swelling, 109 intermolecular, 32 interstitial substance, 80, 8 1 intrability, 197 intramicellar, 80 intramicellar processes, 81 intramicellar swelling, no intramolecular, 32 intrinsic birefringence = i. double refraction, 88, 298, 308 intussusception, 285, 306, 307 ion lattice, 28 ion radii, 148 ion series, 148 ionization, 232 isoelectric s. I.E.?. junctions, 67, 145, 159 theory of, 145, 184 Kappenplasmolyse, 197 karyokinesis, 238, 242 karyolymph, 215, 217 keratin, 338 keratin- myosin group, 365 kinetochore, 224 kinoplasm, 181, 200 konyaku, 322 lamellar birefringence = layer b., 255 lamellar structure, 255, 256 lamellar systems, crossed, 306 lattice, 26 chain, 34, 36, 69 crystal, 26, 69, 97 heteropolar, 28 homopolar, 28 ion, 28 layer, 36 primary valency, 28 molecule, 30, 35 lattice arrangement, 12 lattice plane, 26 lattice regions, 78 layer birefringence, 255 layer composite body, 82 layer lattice, 36 layer structure, 255 lecithin, 55, 138, 252 leptonema, 234, 237 leptones, 79 light absorption, 252 lipid filter theory, 197, 267 lipid theory, 197 lipidic drops, 193 lipids, 137, 267 lipophilic groups, 47 lipophily, 44 liquid crystals, 5 1 lintnerization, 317 long-range forces, 158, 177 long-range spacings, 367 lyo-enzyme, 208 macrocoacervation, 20 macromolecular chemistry, 5, 59 macromolecules, 59, 81, 125, 160 main chain spacing, 365 main valency forces, 31 maltose, 39, 314 mannan, 59, 321 mannose, 60 matrix, 181, 226 medullary sheath, 362 meristematic cell walls, 279 meristematic cells, 281 mesophases, 51 4IO SUBJECT INDEX methylene bridge, 155 micellar, 79 micellar framework, 75, 81 micellar phase, 79 micellar strands, 73, 96 micellar structure, 77, 81, 105 micellar theory, 76 micellar texture, 81, 94 micelle, 76, 81 microcoacervation, 20 microfibrils, 104 microscope, centrifuge, 194 electron, 3, 116 microscopy, polarization, 82 microsomes, 172 Mischkorper, 83 miscibility, diagram of, 46 mitochondria, 173 molar cohesion, 32 molecular framework, 73, 80, 81, 201, 341 molecule lattice, 30, 35 monolayer, 49 monomolecular films, 49 monosaccharides, 37 morphogenetic configura- tions, 179 morphology, i mosaic growth, 286 mosaic theory, 197 muscle fibres, 352 muscular contraction, 358 mutation rate, 231 myelin forms, 54, 247 myelin sheath, 360 myofibrils, 353, 357 myogen, 352, 353 myosin, 352 nematic state, 52 nerves, 360 neurofibrils, 363 neurokeratin, 361 neurolemma, 364 neuronin, 363 nucleal reaction, Feulgen's, 214 nuclear membrane, 217 nuclear sap, 215, 217 nuclear spindle, 222 nuclear staining, 218 nucleicacids, 23, 212,228, 237 nucleolus, 221 nucleoproteins, 210, 227, 240 nucleotide, 208, 212 nucleus, 210, 242 active, 215 chromosomes, 224 fine-structure, 215 fixed, 216 hereditary processes, 230 molecular constituents, 210 nutrition line, 162 optically negative, 53, 87 optically positive, 53 organ, i organizer, 179 orientation angle, 90 orientation double refrac- tion, 89 ossein, 350 ovalbumin, 93 paraplasm, 181 pasting, 324 Patterson-Fourier diagram, 328 pectins, 60, 293 pectic substances, 60 pentosans, 61 permeability', 197 selective, 201 formula, 204 theories, 199, 201 permutoid reaction, iii PH, 156 phase, 6, 15, 68 meso-, 51 pseudo-, 69 phase boundaries, 18, 40 phase separation, 191, 19-1, 196 phenol, 46 pheron, 236 phosphatides, 138 phospholipids, 265, 266 phosphorolysis, 314 photo-elastic effect, 89 phragmoplast, 224, 281 phyllochlorine complex, 247 physiology, 7 pigments, 248 plasma, active, 181 plasma gel, 186 plasma sol, 1 86 plasmalemma, 199 plasmic strands, 170 plasmolysis, 170 cap-, 197 plasmoptysis, 17c plasmosin, 184 plastid membrane, 256 plastin, 183 platelet composite body = layer composite b., 82 points of attachment, 67 polarity of cytoplasm, 190 polarizability, optical, 92 polarization microscopy, 82 polyarabinan, 61 polygalacturonic acid, 60 polymer uniform sub- stances, 64 polymeric homologous, 63 polymerization, 58 polypeptide chains, 132, 134, 339 polysaccharides, 59 polystyrene, 67 porphyn ring, 248 proline, 133, 346 proplastids, 245 prostethic group, 208 protamines, 211 protein cr^'stalloids, 194, 221, 326 protein fibrils, striated, 368 proteins, 132, 141, 184, 211, 326 fibrillar, 135, 364 globular, 135, 141, 327 reserve, 326 protopectin, 281 protoplasmic flow, 164, 186 protoplasmic fibrils, 194 pseudophase, 69 purine, 213, 214 pyrimidine, 212, 214 ramie fibres, 97, loi, 104 rays, ionizing, 231 red blood corpuscles s. ery- throcytes, 262 redox potential, 156 reserve proteins, 326 resolving power, 1,119 reticular colloids, 74, 81 reticular structures, 72 reticular systems, 70, 73, 81 reticulum, 216 rn, 136 ribonucleic acid, 215 rigidity, modulus of, 171 ring diagram, 99 ring texture, 94 rodlet birefringence, 84, 114, 309, 321, 334 SUBJECT I,NDEX 411 iodlet composite body, 82 rodlet double refraction s. rod let birefringence rubber, 59, 341 sarcolemma, 353 sarcomere, 354 sarcoplasm, 353 scattering, angle of, 90 selective permeability, 201 selectivity constant, 203 sensitizer, 21, 266 sericin, 331 shadowing, 122 short-range order, 107 shrinkage, iii, 149 sickle diagram, 99 side chain spacing, 365 side chains, polypeptide, 133' 135 silk, loi, 531 silk fibroin, 331 smectic state, 52 sol-gel transition, 187 sol solution, 65 sols, 8, 75 solvation layer, 19 spacing, 26 backbone, 336, 365 main chain, 365 side chain, 365 specificity, 134, 372 sperm nuclei, 220, 228 spermatozoa, 274 spherite texture, 86,95, 316, 322 spherites, 86, 95 spindle fibres, 222 spinning capacity, 170 spiral growth, 304 spiral structure, 226 spiral texture, 95, 291, 304, 305 spongioplasm, 181 sporangiophores, 304 sporopoUenin, 297 stabilizer, 266 starch, 59, 310, 318 starch grams, 310, 315, 318 sterines, 138 stimulant substances, 209 Stokes' law, 167 strain theory, 223 stretching experiments, 105 stroma, 244, 256, 264 stromatin, 265, 266 structural chemistry, 4, 24, 33 structural elasticity, 164 structural principles, 24 structural viscosity, 64, 66 structure, 81 crystal, 26 framework, 182, 201 gel, 66 micellar, 77, 81, 105 reticular, 72 spiral, 226 surface, 48 sturine, 21 1 suberin, 297 sugar, 37 sulphur bridges, 155, 340, 342 super-crystalline, 53 surface elasticity, 200 surface energy, 43 surface films, 47, 143 surface growth, 282, 307 surface skin, 41 surface structure, 48 surface tension, 16, 40, 47, 165 Svedberg loile, 125 Svedberg unit, 141 swelling, 55, 109, 148, 346 intermicellar, 109 intramicellar, no limited, 67 syneresis, 75 tanning, 75, 347 target area, 232 target theory, 231 teeth, 350 tendons, 345, 346 tension double refraction, 89 texture, 81 fibre, 94, 290, 303, 332 fibroid, 94, 304, 332 fibrous = fibre foliate, 96, 304 micellar, 81, 94 lamellar, 255, 256 ring, 94 spherite, 86, 95, 316, 322 spiral, 95, 291, 304, 305 tube, 94, 282, 288, 305 tubular = tube thixotropy, 66 thymonucleic acid, 214, 220 tip growth, 282 tonofibrillae, 338 tonoplast, 198, 200, 205 membrane, 205 topochemical reaction, 111 transpiration, cuticular, 301 tube texture = tubular t., 94, 282, 288, 305 tubular texture = tube t. tunicin, 129 turgor extension, 288 turgor pressure, 288 turgor tension, 289 Tyndall scattering, 8 tyrosine, 133, 336 ultracentrifuge, 10, 61, 141 ultrafilter, 127, 199 ultrafilter theor}', 197 ultrafiltration, 8 ultraviolet absorption, 219, 226 ultraviolet dichroism, 229 unit cell = elementary c, 26, 97, 319 unit of assimilation, 261 vacuoles, 23, 191 vacuolization, 22 valency, primary = main v. 28, 31 residual, 147 secondary, 31, 147 valency angle, 37 valency bonds, 29, 146, 153 valency forces, main, 31 valency lattice, primary, 28 valency rule, 1 5 1 Van der Waals cohesive forces, 31, 185 virus, 125, 240 virus protein, 240 viscose, 333 viscosity, 146, 166, 181 structural, 64, 66 vital staining, 192, 218 vitamins, 208, 243, 249 vitrification, 196 wall 'ension, 289 waxes, 137, 292, 295, 297 Wiener's formula, 84 wool, 338 xanthophylls, 246, 249, 250 xerogel, 1 1 1 X-ray analysis, 25, 26 gels, 96, 1 1 3 muscle, 355 X-rays, 25, 113 xylan, 61, 310 Zugfasertheorie, 223 zygonema, 234 PRINTED IN THE NETHERLANDS BY DRUKKERIJ MEIJER N.V., WORMERVEER AND AMSTERDAM 1