BOUGHT WITH THE INCOME OF THE SAGE ENDOWMENT FUND THE GIFT OF 1891 ft.30 4-a3 g ft yX \14. 9306 Cornell University Library arV10478 Researches on cellulose II (1900-1905) 3 1924 031 216 538 olin.anx Cornell University Library The original of tliis bool< is in tlie Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924031216538 RESEARCHES ON CELLULOSE ir BY CROSS AND BEVAN. CELLULOSE An Outline of the Chemistry of the Structural Elements of Plants. With reference to their Natural History and Industrial Uses. With 14 Plates. Crown 8vo. I2i. net. RESEARCHES ON CELLULOSE, I 1895-1900. Crown SVo. 6s. net. ■ ■ • ; i RESEARCHES ON CELLULOSE, II 1900-1905. Crown 8vo. 71. 6d. net. RESEARCHES ON CELLULOSE, III 1905-1910. Crown 8vo. ^s. 6d. LONGMANS, GREEN & CO., 39 Paternoster Row, Loudon ; New York, Bombay and Calcutta. RESEARCHES ON CELLULOSE II (1900-1905) BY CROSS & BEVAN (C. p. CROSS AND E. J. BEVAN) SECOND EDITION LONGMANS, GREEN AND CO. 39 PATERNOSTER ROW, LONDON NEW YORK, BOMBAY AND CALCUTTA 1913 - All rights reserved 5 PREFACE Chemists of previous generations have paid relatively little attention to the amorphous or colloidal state of matter ; and the appearance of uncrystallisable products of reaction, in in- vestigations of carbon compounds, usually evoked the epithet of ' Schmiere,' applied rather as an epitaph. This procedure was perhaps the expression of a constitutional preference for bodies of certain typical physical characteristics, but actually we think it implied the recognition, partly reasoned, partly instinctive, that in the colloidal state we lose touch of the organic relationships which obtain in the great groups of carbon compounds defined and represented by constitutional formulae. If cellulose had been thus met with as an amorphous pro- duct of reaction — that is, in its alternative structureless forms — it might very well have been included in the heterogeneous class of ' Schmier ' compounds ; it certainly would not have attracted the attention of investigators to the degree of devotion of which it has been the object. It might be held that the attrac- tiveness of ' cellulose ' as a subject of investigation lies in con- siderations of utility, and that only its disproportionate technical importance could have supplied the motive for researches which lack the clear direction and control of the leading antecedent generalisations of the science. Obviously, on the other hand, the prominence of cellulose in the arts is the vi Researches on Cellulose, II correlative of its position in the ' natural ' world ; and this, its prominence in the natural order, is the basis of its attractiveness to us as students of natural philosophy. The purpose of this volume is to record the recent advances made in our knowledge of the subject-matter, but more ex- pressly to apply the results to the clearing of the issues con- tained in the question, ' What is cellulose ? ' The investiga- tions of chemists through two generations have contributed a large amount of exact knowledge of the subject-matter, but the central question is unanswered, and there are very few attempts at its solution. During the period 1900-05 there have appeared some valuable contributions to the science of the subject, the more important of which we have reviewed in this work. We have ventured to add some critical notes, which we hope will rank as a criticism of appreciation : the intention being to point the particular contributions to the issues contained in the central question, on which the authors themselves are, for the most part, silent. Our main purpose is to establish, at a juncture which we consider to be critical, as it is favourable to progress, the de- finitely agnostic position, which involves the recognition of current views of the constitution of cellulose as inadequate, reopens the matter from a more comprehensive point of view, and reduces all or any interpretations of the experimental quan- titative facts to their present actual level of ' theoretical ' anti- cipations of final generalisations. In the next place we desire to state the grounds for a considerable widening of the scope ■of the investigations which will be necessary for reaching such final generalisations of inclusive and exclusive value. There have been many attempts to solve the problem of the consti- tution of cellulose in terms of molecular formulae. These Preface vii essays postulate ' molecules ' as the ultimate reactmg units of cellulose, and generally extend the analogies of the ' molecular state ' to bodies of this class without, we think, sufficient in- quiry as to whether such methods of interpretation are really adequate. There are valid grounds for a considerable modi- fication of these methods of interpretation and an extension of the experimental investigations, in order to pass from a sphere of discussion which is rather academic than real, into a more direct and comprehensive attack of the fundamental problem, ' What is cellulose ? ' In dealing with the subject from this point of view, we cer- tainly put scientific before industrial developments, at the same time recognising that these terms rather lose their significance in a region where the industrial developments are entirely scientific in character, and ' theoretical science ' has not suggested any positive directions of progressive investigations. From the present, however, we consider that the lines of in- dustrial progress will be rather directed by the application of general theory than the pursuit of technical objectives, and we have endeavoured to illustrate this conclusion in the indica- tions of investigations ,which are urgently needed for the immediate solution of technical problems, and ultimately as a contribution to the solution of the central problem. We are indebted to our friend Prof. Sir William Ramsay for important suggestions, and we also wish to recognise the valuable assistance of our collaborator, Mr. J. F. Briggs, in the production of this work. 4 New Court, London, W.C. SECOND EDITION We have taken advantage of this Second Edition to make verbal conections; but as the matter is a record of progress as well as of facts it has been judged better not to attempt any revision. May, 1913. Pages CONTENTS SECTION I General risuTtie and forecast in Three Chapters 1. Cellulose as a typical colloid .... i-8 2. Cellulose as a chemical individual . . . 9-19 3. Cellulose and structural forms — Dimensions . 20-29 SECTION II Account of original investigations in period 1900-1905 : abstracts of authors' communications followed by critical examination of matter and results from point of view of theoretical forecast of Section I . 30-160 A. Reactions of Synthesis Nitrie Esters. — Interactions of Cellulose and Nitric Acid — (i) Ex- haustive quantitative statistical investigation — Liinge. (2) The reaction at limiting phases (HNO3 only) — Knecht. (3) Me- chanism of reaction with HjSOj present : Mixed esters — (a) Cross, Bevan, and Jenks ; (b) Hake and Lewis. (4) In- vestigation of stability and definitions of normal instability — ^Will. (5) An investigation of solubilities of cellulose nitrates referred to a theory of constitution — Bemadou. (6) Relations of nitrates and of cellulose to atmospheric moisture — (a) Masson ; (6) Will 30-83 AcetoSulphates. — Interactions of Cellulose and Acetic Anhydride ^in presence of Sulphuric Acid->-(7) Further study of mixed esters — (a) Cross, Bevan, and Briggs ; (6). ibid. . . . 83-93 Researches on Cellulose, 11 Cellulose-Xanthogenic Acid. — Interactions of Cellulose Alkaline Hydrate and Caibon Bisulphide — (8) Investigation of series and analytical methods — Cross and Bevan . Cellulose and Alkaline Hydrates. — (9) Investigation of 'mer- cerisation ' and attendant modifications, structural and chemical, of fibre-substance (cotton) — (a) Pope and Hiibner (i)ibid Theory of Dyeing. — (10) Relationship of various colloids, in organic and organic (including textile fibres), to colouring matters, organic and inorganic — (o) Bilz ; (6) Suida ; (c) Suida ; {d) Theoretical discussion of phenomena — Dreaper Electrolytic Phenomena. — (ii) Investigation of conductibility of cellulose in presence of water — (a) Cross and Bevan ; (6) Notes on cathode effects, communicated by C. Darling ; (f) Note communicated by A. Campbell .... Constitution of Cellulose. — (12) Controversial discussion of con- stitutional formulae — (a) Theoretical paper — Green ; (6) Cri- tical paper — Cross and Bevan ; (c) Rejoinder — Green .^ B. Reactions of Resolution and Decomposition Hydrocellulose. — (13) (a) A study of the interactions of cellulose and mineral acids, and criticisms of conclusions of previous investigators — Stern, (b) Critical discussion of foregoing — Cross and Bevan Pages 93-105 106-116 116-127 IZ7-131 131-X37 138-142 Mixed Esters — Chloracetyl derivatives of Hydrocelluloses, &-C. — (14) A comparative study of starch, glycogen, and cellulose to determine molecular weights of products of resolution — Skraup 143-145 Animal Digestion and Assimilation of Cellulose, — (15) A comparative study of cellulose and typical fodders and food- stuffs reduced to expressions of quantitative efficiency — Kellner 145-151 Destructive Fermentations. — (16) An investigation of bacterial resolutions of cellulose with quantitative statistics of end- products — (a) Omelianski ; {b) Notes communicated by Macfadyen 151-154 General summary and conclusions as to experimental verifications of views expounded in Section I . . 154 Contents xi SECTION in Pages Technical progress in Cellulose industries and general forecast of technological developments from point of view of matter of Section I . . 1 61-174 Additional Bibliographical Notes . . . . 175-180 Index of Authors 181 Index of Subjects 182-184 RESEARCHES ON CELLULOSE II SECTION I CHAPTER I CELLULOSE AS A TYPICAL COLLOID In our previous treatises and special publications we have of necessity written under the influence of the dominant idea or aim of the science, which in relation to cellulose would imply that its reactions and chemical history should find expression in a constitutional formula — a formula sufficiently exact to define its relationship to the great mass of carbon compounds, in the order and arrangement constituting the imposing fabric of ' organic ' chemistry, and sufficiently com- prehensive to include and set out its various re-activities. But at the same time we have been aware that this assump- tion is a pre-judgment, if not a prejudice ; and in view of the insufficiency of the evidence in relation thereto, it was neces- sary to adopt the more ' evolutionary ' plan of methodically recording the experimental facts, without reference to any- thing beyond the logic of the method. The basis of the method was an arrangement of the experimental matter on the broadest lines of classification, namely : II. B 2 Researches on Cellulose, II 1. Reactions of synthesis. 2. Reactions of decomposition, chiefly hydrolysis, acid and alkaline, and oxidation. To study and set out the empirical facts in relation to the prototype of the group — viz. cotton cellulose — was the necessary preliminary to dealing with the entire group. To differentiate and classify the members of the group, we could only take advantage of reactions of decomposition. The celluloses, as a group, have a close general resemblance to the prototype in giving synthetical derivatives of similar character and composition ; but in relation to hydrolysis and the products of decomposition there are divergences sufficiently wide to constitute a positive basis of classification ; notably (i) resistance to hydrolytic agencies, and (2) the proportion of furfural obtained by decomposition with condensing acids. Lastly, the celluloses occur in the natural world not in the pure state, but combined or mixed with groups of divergent character. On these broad lines of difference we are enabled to classify as chemical individuals these structural components of vegetable tissues, and they present three well-marked groups — the pecto-celluloses, ligno-celluloses, and cuto-cellu- loses. Now, from the point of view of systematic chemistry, it might be urged that this diversified group of compounds can or must be treated independently of all other considerations than those appertaining to matter as matter. Questions of visible structure, questions of physiological function, are irrelevant. It is self-evident that each of the great groups of carbon compounds has innumerable points of contact with the natural world, and that there is a ' natural history ' of carbon compounds which complements their systematic interior relationships. But it is true, and perhaps remarkably true, that the latter have entirely excluded the former in the Cellulose as a Typical Colloid 3 evolution of the science ; and more than this, that considera- tions drawn from the vital origin and relationships of ' organic ' compounds have proved rather an impediment than an aid to progress. This aspect of the science has been exhaustively dealt with in the recently published and admirable treatise of Meldola's on ' The Synthesis of Vital Products,' and it would be superfluous on our part to re-argue the theme. ' Organic chemistry ' is itself a structure, of independent and self- contained existence, of which ' life ' is an exterior incident and without modifying influence, so far as we can see, on its fundamental basis — that is, on the properties of matter as matter. But while we accept this substantial conclusion in its latest presentment, we may be permitted to draw issue with the author on the form in which it is presented, since this has an immediate bearing on our subject. The form of the author's ' argument ' is indicated by the title of the work, ' The Synthesis of Vital Products.' But the products dealt with are not 'vital.' They are strictly 'by-products' of vital actions or processes, and these are immediately purely chemical in character. There is an implication in the title that ' chemical ' synthesis — i.e. laboratory methods of production — may be extended to products which are, in the strict sense of the term, vital. While we do not contend that the proposition is unthinkable, we hold that its introduction, even by suggestion, into the sphere of the practical politics of the science, tends rather to shelve and obscure problems which are now pre- sented to us for investigation. These problems concern the relations of organic form or structure to chemical composition, which, in the case of the cellulose group, are evidently very intimate. It would, of course, be possible to ignore this correlation, and to write a treatise on cellulose in which the various members of the group are dealt with as chemical B 2 4 Researches on Cellulose, II individuals purely, and all questions of form or structure merged into their common description as amorphous or colloidal aggregates. We suggest that this would be an arbitrary procedure, if only in view of a priori considerations, notably : First, that the celluloses and allied compounds are formed exclusively as the products of vital processes, of which they are, in turn, the arena; secondly, they are produced in specialised mass- forms ; thirdly, their formation postulates a condition of matter in which the molecular energies are dominated by a physical force majeure the effects of which persist in the underlying ultimate or chemical configuration of the aggregate. On this view, visible and external structure is intimately correlated with the ultimate configuration of the chemical units : becomes a positive subject of investigation. But on a posteriori grounds the special study of the celluloses compels the inclusion in the area of investigation of all questions relating to external form. As in many other developments of the science, industrial applications have pre- ceded scientific investigation, and the newer industrial uses of the celluloses are largely conditioned by the facts : 1 . That the structural features of the celluloses persist in many of their derivative compounds. 2. That the solutions of these derivative compounds are plastic, and are employed in producing artificially moulded solids in continuous masses and any required dimensions, fibres, films, and massive aggregates. It is an obvious consequence that the more ' progressive ' researches with which we shall have to deal in this work are intimately bound up with questions of external form or structure. We do not wish to imply that there is any imme- diate significance to us in any particular form in which the natural celluloses are produced, nor that the artificially pro- Cellulose as a Typical Colloid 5 duced solids are to be regarded as structures in the same sense. They are rather internally homogeneous and structure - less ; solidified in continuous mass, with any required external form and dimensions. But it is none the less true that the physical properties of these artificial forms are definitely related to the structural qualities of the original cellulose, and these again are shown to be definitely related to chemical constitution. We may sum up the argument in the statement, which we shall have occasion to elaborate at a later stage, that cellulose is essentially an aggregate; the nature of this aggregate is a present definite object of investigation. It is by no means necessary that our ideas of discrete constituent molecular units apply to aggregates of this class. It is, as we know, generally stated that cellulose is a complex polymeride of a unit molecule of dimensions Cg, and of the general type of the carbohydrates, aldose or ketose. As a matter of fact, this view has not contributed to the progress of research ; and its vagueness is seen in the necessity for continually altering the dimensions of the supposed polymeride. Thus, in the case of the nitric esters, which have been more extensively investi- gated than any other group of derivatives, it has been found necessary to abandon the Cjj formula (Eder) in favour of the C24 formula (Vieille), in order to avoid the formulation of nitrates representing fractional proportions of the assumed ultimate Cg. And later researches (p. 30) invalidate the assump- tion that the arbitrarily enlarged unit has any final value. If we could assign any final value to the experimental numbers obtainable in this and other series of derivatives, we should be justified in the assertion that cellulose is an aggregate of essentially variable dimensions. This hypothesis is unavoidable as an alternative to be critically examined. It is of course evident that cellulose is a representative 6 Researches on Cellulose, II colloid, and therefore that the investigation of the colloidal state generally, which of late years has taken very definite shape, will have furnished some conclusions available for our special case. We cannot attempt here any exhaustive notice of the recent extensive literature of the subject, and, without preface, we proceed to apply a critical statement contained in a dissertation, 'Neue Gesichtspunkte zur Theorie der Kolloide,' by E. Jordis.^ This author says (p. 63) ; ' In opposition to prevailing opinion I regard the colloids, and more particularly the hydrosols, not as pure forms of matter, but as the analogues of chemical compounds in dilute aqueous solution . . . formed as the products of a normal dissociation and hydrolysis of crystalloids in aqueous solution.' This conclusion follows from ^an examination of the essential conditions presented by typical inorganic colloids, viz. (i) These colloids are, in the main, compounds of poly- valent elements ; hence they break up by dissociation (hydro- lysis) into mixtures of ions of maximum complexity. (2) This complexity is particularly determined by the amphoteric functions of the hydroxides. (3) The limiting equilibrium as between hydrosol and hydrogel is a definite saline equilibrium or neutral point. Thus in the case of sodium silicate the neutral point is reached with the ratio NajO. 308102; and the hydrosol which results from treating the silicate with excess of HCl may be formulated as a species of chloride [Si(OH)J„-Si^ p, " ; one of the OH groups of the hydroxide having a basic function. Similarly in the case of ferric hydrate, the neutral point is formulated as [Fe(0H)3]„-Fe^ „, ''^ ; and the hydrosol is stable up to the limiting ratio iCl : loFe. In the presence of ' Eriangen, February 23, 1904. Cellulose as a Typical Colloid 7 organic acids, &c., stable alkaline hydrosols are formed, the (OH) groups of the hydroxide assuming an acid character. These two cases are cited as types. Similar conclusions have been arrived at independently in regard to cellulose. In describing experiments which demonstrated a specific electrical conductivity of cellulose in presence of water, together with electrolytic reactions observed in a circuit completed by moistened cellulose, we used the expression ' It is another link in the chain of demonstration of those characteristics . . . which, for want of a better term, we describe as "saline"' (J. Chem. Soc. 1895, p. 451). In previous works we have pointed out that the actions of solvents of cellulose can only be explained as the result of its amphoteric activity; that the solution of cellulose in zinc chloride and cuprammonium solutions are to be regarded as containing a species of colloidal double salt. These considerations result in a working hypothesis which may be briefly formulated as follows : Cellulose does not react as a polymeride of pre-formed hexose groups of rigid configuration, but as a labile complex of groups of varying dimensions representing a state of matter somewhat analogous to that of a solution of a saline electrolyte — that is, it reacts rather as a solution-aggregate than by a succession of molecular combinations ; the masses actually reacting following the stoichiometrical ratios proper to the dimensions of these ultimate groups, and retaining their relationship in the aggregate, which is thus progressively modified by the entrance of the new groups. This view, implying a species of continuous reactivity, may appear at first sight to conflict with current teaching ; but on further examination it will be found that the paradox on the surface involves no antagonism in essential principle. We may still regard cellulose as substantially a polymeride of 8 Researches on Cellulose, II hexose groups, but modifying our view of such an aggregate as of invariable function and fixed dimensions. We take the constituent molecules as so related that when reaction takes place, more especially under conditions which involve the presence of hydroxyl or of hydrogen ions, the effect of re- action upon a particular molecular unit is communicated to other units, which are consequently modified in function, becoming more or less basic or acidic ; hence a complex reacting unit of variable or changing function. It is also necessary to assume that cellulose is ionised both as acid and base, and is then amphoteric. We have thus at once (a) an acid, {p) a base — that is, hydrogen and hydroxyl ions, and the corresponding complex residual groups — and {c) non-ionised molecules, susceptible of similar transformations, with the changing conditions of equilibrium ; hence a reacting unit of variable dimensions. Lastly, the exceptional structural rela- tionships of cellulose and its derivatives warrant the assumption that the complex (c) may act as a solvent of the ionised groups [a) and (b). In all reactions of cellulose, as cellulose, the aggregate is maintained, and with it the associated colloidal character, which also implies structural integrity in the sense in which this has been defined above. Reactions which break down the aggregate are attended by the formation of end-products of decomposition of molecular character and dimensions; but these represent a new equilibrium, and are not to be regarded as pre-formed in the cellulose aggregate. The cellulose aggregate is modified by the conditions under which it reacts ; just as every colloidal salt in solution would be affected by every change determined in the solution, such as of temperature, dilution or concentration, the addition of ' neutral ' compounds, and, of course, more profoundly, by active reagents combining specifically with its constituent groups of ions. CHAPTER II CELLULOSE AS A CHEMICAL INDIVIDUAL The discussion of the previous chapter will have suggested that our subject involves much 'that is undetermined if not vague or nebulous. It will be the purpose of subsequent ■chapters to deal with recent investigations as contributing to the elucidation of those problematical issues involved in general questions of constitution, and the actual state of matter composing the cellulose complex. They may not have been undertaken with this object immediately in view ; in fact the authors' purpose is, in many cases, rather to reduce some particular phenomenon to the basis of exact quantitative expression, leaving out of consideration the issues which are vague and undetermined. But, with this empirical aim, there is a factor at work leading the authors to explore simultaneously, and with more or less set purpose, the region of theory. Now it will be generally recognised that the chemistry of cellulose and its derivatives presents a large accumulation of experi- mental matter, of which the quantitative relations are as exact and well grounded as those of any other section of the science ; notwithstanding the fact that we are dealing with essentially uncrystallisable complexes, colloidal aggregates, in fact, of which therefore we have not the associated criteria of 'purity,' and constitutional relationships, belonging to bodies which can be examined in the ' molecular ' state. It will facilitate the object we have in view to separate such empirical data as can be briefly set out from those lo Researches on Cellulose, II which involve the undetermined elements of our subject, and which require discussion from the point of view of alternative hypothesis. We shall therefore in this chapter give a statement of the chemical characteristics of cellulose, which may be taken as a condensed summary of the experimental facts, such as might be included in a systematic text-book of the science : Cellulose. — Generally the non-nitrogenous skeleton of vegetable tissues. Type : the fibre-substance of cotton, puri- fied from associated 'impurities' by processes of (i) alkaline hydrolysis and oxidation (bleaching) ; (2) of digestion with hydrofluoric acid, &c., to remove mineral impurity. fC 44'4) Composition. — Elementary j H 6-2 I, whence the empirical iO 49-4) formula CgHjoOj. Constitution undetermined. — Is variously regarded as : (r) Polyhexose (anhydride) 0'Oi6-o"05o Jute . I-5-4-0 „ O'O20-O-O25 Coaise textiles Sisal . i'5-6-o „ o"oi5-o"026 and H Phormium 5-0-I5-0 „ O'0IO-0*020 rope-making Pinewood ) I-0-2-0 „ O*OI5-0"020 I, (Tracheids) |- Paper-making Esparto . J O-5-3-0 „ o-oio-o'oi8 Only in the exceptional case of cotton is the 'ultimate, fibre ' the naturally occurring form of the fibre substance. The general case is that of a fibre bundle, simple in the case of flax, hemp ; more complex in the case of jute and the fibro-vascular bundles of monocotyledons ; and a hetero- geneous aggregate in the case of esparto and many of the woods. With the more complex structure the celluloses are associated with various chemical groups (compound cellu- Cellulose and Structural Forms — Dimensions 21 loses), which are attacked and removed by the various treat- ments of hydrolysis and oxidation by which the celluloses are isolated. For our present purpose we are concerned only with the celluloses in the form of ultimate fibres. These are the unit elements of structure of the yarns and threads which are the basis of textile fabrics ; and the mechanical properties of these, as well as the processes by which they are mechanically prepared and ultimately spun, are obviously determined by their simpler elements of form — that is, their dimensions. The mechanical principles involved in the production of fine textile yarns from these discontinuous units are, first, the reduction of these to a continuous untwisted sliver, in which they are parallelised ; secondly, the drawing and twisting of the fibres composing the sliver in continuous length. The tensile pro- perties of the resulting yarn depend mainly upon the twist, partly upon the adhesion of the more or less closely spun fibres. There are numerous variations of the process, such as the wet spinning applied to the bast fibres and notably flax, the passage of the sliver through a bath of warm water facilitating the ultimate subdivision of the bundles of fibres. In the coarser textiles, such as jute, it is evident that the spinning unit is an aggregate or bundle of the ultimate fibres, which are too short (2 to 3 mm.) to admit of manipulation. We are only concerned with this general and superficial " view of spinning processes in passing to the main subject of this chapter, which is the tensile characteristics of the ' arti- ficial silks ' or lustracelluloses. In these we have the common feature of a solution of cellulose, i.e. of a cellulose compound, drawn at uniform speed through a fine orifice and solidified in continuous length. The resulting thread may be regarded as structureless and of more or less regular dimensions. The mechanical properties of such a thread — resistance to strain, elasticity, and extensibility — ^are therefore properties of the 22 Researches on Cellulose, 11 cellulose substance independently of grosser structural details such as characterise the ' natural ' cellulose fibres. Moreover, the regularity of form gives a particular precision to the measurements of tensile properties. We are thus able to for- mulate a normal basis of investigation from which the whole of the physical properties of the celluloses may be inquired into with some prospect of revealing the relationships of ex- ternal or visible form to the invisible units or aggregates of which the fibre substance is made up. A few illustrations will indicate the directions of inquiry. The ' artificial silks ' are classified for relative fineness on the basis of the metrical weight-length units which are used in the silk industry. The 'denier' (metrical) represents the weight in milligrammes of the unit length, viz. lo metres. It applies equally to the single filament (brin) or to the assemblage of these — usually from lo to 20, which make up the weaving unit. In speaking therefore of a yarn of 80 deniers having a 'tenacity' of 96 grms. and an 'elasticity' of 12, we are shortly expressing the 1 ultimate facts as follows : The yarn weighs o'oSo grm. per lo-metre length, and breaks when stretched with a weight of 96 grms., having undergone an elongation of 12 p.ct. in length. Assuming 16 unit fila- ments in the yam, each is of 5 deniers and would bear -^ of the stretching weight. Taking an average diameter of the unit at o'o3 mm., this implies a sectional area of 0-000707 mm.D ; and the tenacity may therefore be expressed as being equal to = 8,486 grms. per mm.D, ©•000707 which at once conveys the relationship of cellulose to other structural materials which are capable of being produced as a regular solid in continuous length, e.g. the metals. Another form of expression which has advantages in comparing structural quality is the breaking length — that is, the Cellulose and Structural Forms — Dimensions 23 length of the material which,' supposed extended in space, will determine fracture ; in the above case 96,000 X 10 83 12,000 metres. We may compare this figure with the average numbers for the standard cellulose textile yarns taken from a recent work by a leading authority, ' Papierstoffgarne,' Prof. E. Pfuhl, Riga, 1904. Cotton yarns . Ramie Flax ('line') (wet spun) Flax yarns (dry spun) Jute yarns Lustracellulose {'^^^,, Breaking length km. 13-14 II-12 I2-4-I9-3 Ii'8-I2'4 976 12 8'o-i4'o Extensibility p.ct. 3-97 o7g-i-gs 1-1-178 2-5-3 -67 2-0 10 7-18 Apart from the practical significance of these comparisons, they supply the basis for an analytical estimate of the various factors which contribute to structural quality. In the case of the natural fibres these additional factors are the twist of the yarn and the apposition of the surfaces of the component fibres, which, again, will be effected by their minute structure. These questions graduate, in another direction, into the problem of the actual structure of papers and the fibre agglo- merates produced by the paper-making process. A paper is a continuous web composed of discontinuous fibres, which re- quire to be milled or brought by mechanical treatment within a certain somewhat narrow range of lengths, viz. i to 3 mm. The strength of the dried web or agglomerate depends upon the interlacing of the fibrous units, and the adhesion of their surfaces, which latter factor is largely influenced by the colloidal properties of the fibre substance, as well as by the co-operative action of structureless hydrated colloids, either derived from 24 Researches on Cellulose, II the fibre substance or added in the form of sizing agents. The mechanical properties of papers are expressed in the same terms as those of the textile yarns, and it is interesting to compare the range of numbers for the highest classes of papers : — Breaking length Elongation Rag papers, manila papers, and wood-pulp papers (air dry) metres 4,000-8,000 p.ct. 3-8 It is evident from these numbers that the aggregation of shorter units by the paper-making process produces a texture approximating in mechanical properties to the lower grades of spun yarns. It is only necessary to change the form or dimensions of the aggregate from the plane to the cylindrical to produce a yarn available for weaving. An industry based upon these principles is being actively developed, a full ac- count of which will be found in the recent work of Professor E. Pfuhl, ' Papierstoffgarne,' which we have previously cited. In addition to the conversion by a rolling process of the flat strips into the cylindrical form, these are subjected to a twisting process while still in a moist condition, with the effect of increasing the solidity and resistance of the agglomerate. This is illustrated by the following comparative numbers ; — Breaking length Elongation Plane strips — dried Rolled into cylindrical form Rolled and twisted km. 4-17 5-i8 6-41 p.ct. 2 '84 2-71 3-06 Tt is perhaps scarcely necessary to state that yarns of this class show a very large reduction in tenacity when wetted with water, and their range of application is proportionately limited. The same want of resistance, though to a less extent, also Cellulose and Structural Forms — Dimensions 25 limits the usefulness of the lustracelluloses. We are, however, not concerned at the moment with industrial questions, but with ultimate structural phenomena. The action of water in the two cases, though similar in effect, is obviously referable to different proximate causes. The action on the fibre- aggregate is to break up the adhesion of the fibre surfaces by the penetration or imbibition of the liquid ; but the cellulose of the ' artificial silks ' is affected by actual combination with water or hydration, which may be regarded as an incipient solution. Of the ultimate phenomena which are concerned we can only, at present, form vague and conjectural pictures. But the subject will not remain in this condition of indefinite- ness, for the special properties of the lustracelluloses, and generally of the artificial structureless forms of cellulose, supply a new basis for the accumulation of exact measure- ments. This is especially the case in regard to the entire range of hydration phenomena, in which the incipient stages are connoted by the expression 'hygroscopic moisture,' or the corresponding French term ' humidite ' — that is, the property of taking up from the atmosphere and retaining a proportion of water. The proportion of water which the celluloses retain in the air-dry condition is not only a ' constitutional ' fact, but an important factor of their mechanical properties. Some observers, who have investigated these phenomena from a more special point of view, have arrived at different con- clusions; and we shall have occasion to note (see p. 71) more particularly the researches of Orme Masson ' On the Wetting of Cotton by Water and by Water Vapour' (Proc. R.S. 1904, p. 230). From a close study of the associated thermal effects Masson concludes ' that the heat produced by the absorption is of about the same magnitude as the heat of liquefaction of the same quantity of water ' ; and, after discussing the 26 Researches on Cellulose, II bearings of this central fact, ' the view that the water becomes chemically combined to form definite hydrates of cellulose may be quickly dismissed, as there are no facts to support it.' We shall deal subsequently (p. 8i) with the evidence which Masson regards as non-existent ; but at the moment we are concerned with an incidental discussion which follows the above statement. The author has to deal with the supposition that the moisture is deposited and held as a surface-film. He arrives at an approximate estimate of the actual surface of a given mass of cotton ' on the assumption that the fibres are uniform cylinders of such length that the ends are negligible,' in which case the surface or area is 4 ze/ / I'Szs x 188 x io~*, in which expression w is the weight, i"S2S the sp.gr. of the cotton substance, and 188x10"* cm. the average diameter of the fibre. To those who are familiar with the minute structure of the cotton fibre — its large and irregular variation from the cylin- drical form, and the complicated details of its imbricated interior texture — it is evident that the assumption is subject to errors so large as to invalidate any calculation of volume or surface. It is true that the author only employs it in disposing of an incidental point, and is evidently aware of its uncer- tainty. But we may, without these elements of uncertainty, apply the calculation to the artificial cellulose tissues, more particularly to the lustracellulose obtained from aqueous solutions of cellulose, since these are of approximately regular form. Thus, taking i"5oo as the sp.gr. of the air-dry product, D the diameter of the unit thread, w the weight of a given mass : • The surfaces of a given length, L, is t D L and the weight w =■ L X i'5 4 so that the surface s, of a given weight w, is w D X 1-5 Cellulose and Structural Forms — Dimensions 27 From these elementary facts and relations an extensive range of investigation is opened out. In the first place, the two obvious factors of the combina- tion of the celluloses with atmospheric moisture can be separately varied. The water must be taken up at the limiting surface, and the rate of combination must vary in proportion to surface. But the total complement of ' moisture ' represents a property of the fibre-substance ; the substance is penetrated by the water molecules, and presumably by osmotic diffusion. This is fully recognised by Masson (p. 252, loc. cit.), and the combination, he says, ' may be regarded as a solid solution of cellulose and water.' With an irregular solid such as cotton, and no measurement of surface of which the errors can be even approximately quantified, the investiga- tion of the simplest of such phenomena is rendered much more difficult. The methods and results described in this interesting paper must therefore be complemented by obser- vations upon the regular forms of the ' artificial cellulose ' — fibres, films, and larger masses — and the results must cer- tainly tend to a closer knowledge of the mechanism of the reactions. But we are not limited to observations upon cellulose ; we may extend them to the compounds of cellulose such as can be brought into homogeneous solution ; notably to the nitrates and acetates. Reverting for a moment to the discussion of the earlier part of this chapter, we pointed out that the more purely mechanical properties are reducible to very exact expression in the case of homogeneous structureless solids producible in regular form, and of dimensions varied at will. It is evident that a study of derivatives which can further be varied at will, over a wide range of compound forms, must contribute to a knowledge of the, as yet, obscure conditions of the solid state. It might be supposed that chemical con- 28 Researches on Cellulose, II Btitution is not causally related to structural properties ; and 1 superficial view of these forms of cellulose and derivatives of the closest external resemblance and of similar structural properties, would appear to confirm the impression. But, as chemists, we cannot limit our views to the proximate and visible forms of matter. We can only think in terrtis of ultimate constituent groups ; and we must put the question whether these are possibly related to the visible forms of the apparently homogeneous matter? In the case of the cellulose group we are certainly able to affirm that structural modifica- tions accompany definite chemical changes in the cellulose complex ; and such modifications persist in the derivative com- pounds obtained from such modified celluloses. Admitting that these experimental facts introduce structural relation- ships, we must pass to the a priori generalisation, upon which basis we are confronted by the following problems : (i) Are the structural properties of the celluloses definitely related to their ultimate configuration, in the chemical sense? (2) If so, is the relationship dominated by the arrangement of the carbon atoms, or are the hydroxyl groups more particularly involved? or (3) Is the causal connection to be found in the superior grades of aggregation or proximate constituent com- plexes of unascertained dimensions ? The evidence at present available points to the latter con- clusion, but also indicates that the hydroxyl groups are not essentially involved, since the esters have generally the same colloidal and structural properties as the parent celluloses. Lastly, we find that with the colloidal properties of cellulose, while there is much in common with the colloids generally, there are specific characteristics which may be assumed to be correlated with the grouping of the carbon atoms. We shall deal with the evidence which has so far accumu- lated, in later sections. The purpose of this chapter is limited Cellulose and Structural Forms — Dimensions 29 to the broad statement of a case for the pursuit of problems of structure, from the grosser limits, through the range of minuter dimensions requiring microscopic measurements, to the ultimate details of the invisible structural units. We have no wish to anticipate the results of investigations, nor even to forecast in definite terms the lines of work which, in our opinion, are best calculated to contribute to the evolution of the subject. SECTION II We proceed to the contributions of experimental matter, of which we shall give an account in the form of abstracts, retaining the several authors' particular views in detailing the substance of their communications, following with a critical discussion of their results from the standpoint of the views which we have set out in the earlier chapters. A. Synthetical Derivatives — Hydroxyl Reactions: Esters. Of the esters of cellulose the most important are the nitrates ; and the following are the more important contribu- tions to the fundamental chemistry of these derivatives : RESEARCHES ON NITROCELLULOSE. G. Lunge (J. Amer. Chem. Soc. 1901, 23, 527). This paper embodies the results of extensive investigations during the period 1897-1900, in which the author has had the co-operation of Weintraub and Bebie (Ztsch. angew. Chem. 1899, 441-467 et seq.). The object of the investigations ' was to study the con- ditions under which nitrocelluloses of various' compositions and various properties can be regularly and with certainty Synthetical Derivatives — Hydroxyl Reactions : Esters 3 1 prepared, and to establish both the composition and the properties of the products obtained under varying circum- stances with greater accuracy than has been done up to this time, as witnessed by the considerable discrepancies among the statements of different authorities.' The form of cellulose nitrated was the so-called ' chemically pure surgical cotton-wool,' further purified by boiling with a weak solution of sodium carbonate and, after thorough wash- ing, extracting with alcohol and ether. The preparation was found to retain o-o6 p.ct. of ash constituents. The details of control of the original reaction, of purification of the products, and of the analytical methods employed are fully described in the two first sections of ,the paper. These are for the most part based upon the standard methods, with such variations and refinements as were necessary to ensure the highest degree of uniformity and accuracy. The author's method for the ' estimation of unchanged cellulose in nitrocelluloses ' may be mentioned at greater length. It is based on the observation that sodium ethylate, which has no effect upon cellulose, 'decomposes the nitro- celluloses almost instantaneously, forming a reddish-brown substance soluble in water.' The reagent is prepared by dis- solving 2 to 3 grms. of sodium in 1 00 c.c. ethyl alcohol (95 p.ct.), and diluting with 100 c.c. acetone. Five grms. of the nitro- cellulose are digested with the reagent at the ordinary tempera- ture for some hours, or heated at 40° to 50° for 20 to 30 minutes. The residue is washed with alcohol and then with water, and finally dried and weighed. It is noted that the method is not applicable to nitrates of lower nitrogen percentage, owing to variations in the results, amounting to 5 or even 10 p.ct. The section reviewing the history of the ' methods for obtaining the various nitrocelluloses ' deals with the funda- 32 Researches on Cellulose, 11 mental question of the actual reacting unit of cellulose, in the following terms : ' The first chemists who worked on nitrocelluloses adopted for cellulose the formula CeHigOg, and consequently spoke of " trinitrocellulose," C6H7(N02)305, as the highest obtainable term, while "dinitrocellulose," C5H3(N02)205, was supposed to represent the composition of " soluble " nitrocellulose. The former expression is still sometimes used to denote gun-cotton (pyroxylin), the latter for denoting collodion cotton. ' Of course it has been known for a long time past that the molecule of cellulose must be a multiple of CgHjoOs, but its real magnitude is not known and is not of much importance in this connection. All we need is to fix the lowest figure by which we can represent the various products obtainable by introducing NOj groups (nitric acid radicals) into the molecule of cellulose. ' Eder ^ obtained four degrees of nitration from the highest downwards, and consequently doubled the ancient formula, as he required the molecule C12H20O10 to represent all his pro- ducts. Vieille,^ however, obtained within the same limits a greater number of distinct products, and found himself com- pelled to use the molecular formula Cj^H^^Ojo for cellulose. This would cause the highest degree of nitration to be called " dodecanitrocellulose," in lieu of the old term " trinitrocellulose " or Eder's " hexanitrocellulose," but Vieille could not obtain this highest term at all and declared "endecanitrocellulose,'' C24H2g(N02)ii02o, to be the highest product of nitration obtainable. The table which follows gives the formulae and nitrogen percentages of the various stages of nitration denoted according to the C24 . . . molecule, which we adopt through- out this paper as being sufficient for representing our present stage of knowledge. '^Ber. d. Chem. Ges. 13, 169. ^Compt. Rend. 95, 132. Synthetical Derivatives — Hydroxyl Reactions : Esters 33 Degrees of nitration Formula Nitrogen NO per p.ct. I gram c.c. Dodecanitrocellulose ( = old C„Hj30,o(NO,)„ I4-I6 225-53 trinitrocellulose) Endecanitiocellulose CmH2902„(N02)u 13-50 215-32 Decanitrocellulose . Cs4H3„02„(NO,)i„ 1278 203-87 Enneanitrocellulose Ca4H3iOa„(N02), 11-98 191-08 Octonitrocellulose ( = old dini- Ca4H3202„(N02)8 II-13 177-52 tiocellulose) Heptanitrocellulose C,,H330,„(NO,), lo-ig 162-53 Hexanitrocellulose CmH3,02o(NO,)5 9-17 146-26 Pentanitrocellulose C,,H3Ao(N03), 8-0+ 128-24 Tetranitrocellulose ( = old mo- C,,H3Ao(NO,), 6-77 108-01 nonitrocellulose) 'Mendelejeffi obtained a product containing 12-44 P-ct. nitrogen entirely soluble in ether-alcohol, whose nitrogen percentage is about midway between Vieille's deca- and ennea- nitrocellulose. Vieille had stated that deca- was insoluble, ennea- soluble, and therefore Mendelejeff considered his soluble intermediate product as a new, individual step of nitration which would compel the adoption of a formula con- taining €48- But as we shall see later on, there exists a soluble decanitrocellulose, and there is hence no reason whatever for making the solubility of Mendelejeff s products a criterion of its nature as an individual compound ; it may very well have been a mechanical mixture of the deca- and ennea- body. I do not, therefore, at present see any reason why we should write our formulas on the basis of €43 . . . and I shall retain the basis of C24 - • . by which all our results can be satisfactorily expressed. Of course, the real formula must be «C24 . . . , but we may neglect the unknown value of n as immaterial for our purposes. ' The lower members of the series of nitrated celluloses, up to deca-, can be obtained by means of nitric acid alone, but as even these in actual practice are always prepared by mixtures 1 Moniteur Scientifique, 1897, p. 510. II. ° 34 Researches on Cellulose, 11 of sulphuric and nitric acids, we have almost exclusively worked with such, and we shall now proceed to discuss the conditions for obtaining the various steps of nitration.' The experimental results recorded in this communication are especially valuable for the internal evidence of accuracy, and may be accepted as establishing a number of constants of importance on the highest stage of nitration obtained from cellulose. By the action of mixtures of the anhydrous acids the maximum stage attained was the endecanitrocellulose (N = 13-5 p.ct.) confirming Vieille (Comp. Rend. 95, 132), and Vignon {ibid. June 6, 1898). On the other hand, the authors confirm Hoitsema's results obtained with mixtures of nitrogen pentoxide and phosphorus pentoxide, viz. the higher stages of nitration (N = 13 '88, 13 '90) approximating to the dodecanitro- cellulose (Ztsch. angew. Chem. 1898, 173). But these higher nitrates are obtainable by varying the composition of the ordinary mixture in the direction of hydra- tion : moreover, as direct products and without fractionation by ether-alcohol to remove lower nitrates. The following table of experimental results establishes this point : No. Composition of acid mixture Nitrogen in product, p.ct. Yield on cotton, p.ct. H.JSO4 HNO3 H^O I 2 3 4 5 6 7 8 6o-oo 62-10 62-95 6372 64-56 68-02 64-55 63-35 27-43 25-79 24-95 25-31 24-65 25-28 25-55 25-31 12-57 12-11 12-10 10-97 10-79 5-70 8-88 11-34 13-62 13-75 13-83 13-75 13-71 , 13-76. 13-72 13-92 173 174 175 17s 175 173 173 It was proved, however, that these nitrates are unstable and readily lose nitric residues until the stage of the (N 02)11 Synthetical Derivatives — Hydroxyl Reactions : Esters 3 5 product is reached. The endecanitrocellulose is, in effect, the highest stable degree of nitration. In reference to the production of this nitrate, it is im- portant to notice that the result is not influenced by a com- paratively wide range of variations of composition of the mixture H2SO4 : HNO3 : Hp. Whereas in the series above recorded the ratio HjSO^ : HNO3 has been maintained at 2-3 : i, this is varied in the following series from 3*3 to 2:1, without influence on the composition of the nitrate : No. Composition of acid mixture Nitrogen in nitro- cellulose, p.ct. Yield, p.ct. HaSOi HNOs HsO 9 10 II 12 13 14 15 16 »7 18 75-33 74-16 72-97 69-90 68-31 67-43 67-32 65-41 63-75 70-68 22-8o 22-12 21-63 20-45 20-49 ig-57 32-53 31-34 30-80 29-31 x-87 3-72 5-40 9-65 11-20 13-20 0-15 3-25 5-45 10-01 13-53 13-51 13-57 13-64 13-61 13-25 13-62 13-57 13-63 13-68 176 175 177 176 172 176-5 175 176 176 It is noted that in the researches of previous investigators it has not been sufficiently recognised that there are three variables in the nitration mixture to be taken into account — the ratio HNO3 • HjSO^, the ratio of both to water, and the presence of lower oxides of nitrogen. The neglect of one or other of these variables is an obvious defect in the investiga- tions of Vieille and of Braby (Mem. des poudres et Salp^tres, 2, 217; 8, III). Maintaining the ratio HNO3 : H^SO^ approximately constant, the effect of varying only the proportion of water D 2 36 Researches on Cellulose, II is seen in the records of the following series of determina- tions : No. I 2 3 4 5 6 7 8 9 lo II 12 Nitrogen, p.cl. Soluble in cther- alcoliol, p.ct. Yield cotton, p.ct. Yield* (Calc.) Acid mixture in p.ct. HaSOi HNO3 H2O 13-65 I3-2I 12-76 12-58 12-31 12-05 "■59 10-93 9-76 9-31 8-40 6-50 1-50 5-40 22-00 60-00 99-14 99-84 IOO-02 99-82 74-22 i"i5 o-6i 173 177-5 176-2 167-0 159-0 I53"0 156-S 144-2 146-0 138-9 131-2 178 175 168 165 163 160 45-31 42-61 41-03 40-66 40-14 39-45 38-95 38-43 37-20 36-72 35-87 34-41 49-07 46-01 44-45 43-85 43-25 42-73 42-15 41-31 40-30 39-78 38-83 37-17 5-62 11-38 14-52 15-49 i6-6i 17-82 18-90 20-26 22-50 23-50 25-30 28-42 * The numbers calculated for yields are added to the author's tables. It may be specially noted that, as the proportion of water reaches 16 p.ct. of the mixture, there is a marked increase in its effect, and ' the group of entirely soluble (ether-alcohol) nitrocelluloses, with from 12-3 to 10-9 N p.ct., is obtained by acid mixtures ranging between 16-6 and 20-3 p.ct. water.' It is clear also that the effects attributable to small variations in the percentage of water must be taken into account in refer- ence to the water formed in the reaction, together with the abstraction of nitric acid. It is also evident that the propor- tion of acid mixture to the cotton treated is an essential factor of the process. An important practical result of the above series is that the conditions intermediate between those of Nos, 7 and 8 are those for obtaining the ' pure octonitrocellulose ' with 11-13 P-ct- N entirely soluble in ether-alcohol. ' The morphological structure of the products is also strongly altered by an increase of water in the acid mixtures. Up to Synthetical Derivatives — Hydroxyl Reactions : Esters 37 15 p.ct. water the original structure is preserved. At 18 p.ct. water the fibres appear somewhat contracted, and the characteristic twist of the cotton fibre is lost. With more water the fibres appear to be torn open, and disaggregate into small particles, which become globular matter. This destruc- tive action is at its maximum with 23 to 25 p.ct. water; by more dilute acids the fibrous structure is much less affected, but here also by prolonged action the fibres are split up into smaller elements.' The influence of temperature is seen in the subjoined tables of results, the first relating to the production of the higher nitrates by treating cotton (2-5 grm.) with fifty times its weight of a mixture of the acids in the proportion of i part HNO3 (1-52 sp.gr.) and 3 parts HjSO^ (f84 sp.gr.). The tempera- ture was progressively varied, and the effect of the time-factor was also studied : Temperature Nitrogen, p.ct. Yield (Gale.) Yield, p.ct. Loss of cellulose, p.ct. No. I 0° 1 i J?°"' \ 7 hours 1071 153 152-29 trace 13-19 174 173-29 trace 2 10° 7 hours I3'37 176 175-78 3 15° 7 hours 13-38 176 175-61 4 TftO \ hour 1272 170 166-14 — 5 19 - 7 hours 13-39 176 175-56 — 6 ArP - \ hour 13-07 173 172-32 trace 7 40 7 hours 13-06 173 169-62 i-6i 8 60° J hour 13-08 173 169-18 1-95 9 4* hours 13-07 173 162-05 5-67 10 J hour 13-07 173 161-23 6-52 II 80° ■ J hour 13-12 173 125-17 27-45 12 3 hours 13-12 173 81-52 52-76 13 The influence of increase of temperature is seen in in- creased rapidity of action; at the higher temperatures the yield progressively falls, which effect does not appear to be due to a direct conversion of the cellulose into soluble pro- ducts, but to a solution of nitrocellulose previously formed. 38 Researches on Cellulose, II The structure of the product is also progressively modified: that obtained at 60° to 80° is a finely fibrous powder after drying. The time-factor was more closely studied in the following series, the conditions of action being those of the preceding series, and the temperature 32°. Time of action, minutes 5 15 30 60 120 Nitrogen, p.ct. 13-27 13-44 I3'47 I3'5o 13 '40 The observations were extended to the series of collodion cottons, using an acid mixture containing HNO3 42-15 p.ct., H2SO4 38-95 p.ct., HjO 18-90 p.ct., the results being as follows : No. Tem- perature Time of nitration, hours Nitrogen, p.ct. Solubility in ether-alcohol, p.ct. Yield, p.ct. Calc. 14 15 16 17 18 17° 17° 40° 60° 60° 4 24 4 11-50 11-59 11-49 11-46 10-81 95-60 99-81 99-58 99-71 99-84 155-1 156-2 148- 1 146-7 152-0 160 160 IS9 159 IS4 The influence of rise of temperature on the yield of pro- duct, and also upon its structural characteristics, is more pronounced. On the latter point the author observes ' the change of structure seems to have ulterior consequences, for, as I am informed upon reliable practical authority, the collo- dion cotton made at such high temperatures does not yield vegetable fibre fit for spinning. Probably, simultaneously with the change of morphological structure the cellulose molecule is also changed.' The effects of varying the proportion of sulphuric acid to nitric acid have been the subject of an extended series of Synthetical Derivatives — Hydroxy I Reactions: Esters 39 experiments, the general conclusions from which are that an optimum lies within the range HjSO^ : HN03 = 0-25-3 -oo : i- Outside this range the degree of nitration is adversely affected, with progressive and proportionate structural modifications of the fibre. When the ratio H2SO4 : HNO3 exceeds 7:1 the fibres are destroyed, and on drying pass into a finely fibrous powder. The quantitative results are set forth in the sub- joined tables : Table I. One-half hour Twenty-four hours Three days | Nitric acid Sul- phuric acid Nitrogen, Yield, Nitrogen, Yield, Nitrogen, Yield, p.ct. p.ct. p.ct. p.ct. p.ct. p.ct. 12-58 162-75 12-62 163-32 ^\ — — 12-66 i65*o2 — — \ 13-45 175-69 13-44 175-77 — — k — — 13-42 175-22 — — I 13-36 174-56 13-39 174-75 — — 2 13-23 174-14 13-32 175-98 — — 3 12-72 i66-i4 13-40 176-44 13-38 175-55 4 — — 13-20 175-12 — — 5 8-14 130-88 unchanged cellulose) 13-10 166-60 TAbLE II. a u ■3 ■a Three days Eight days Fifteen days •1 Ed ■so- IIS B ft z it •0 ll Z 4^. I 6 7 12-63 10-86 162-40 151-60 0-85 10-46 — — z 12-74 i6g-8o none Thirty days. I 8 10-88 144-62 10-7 11-70 152-0 4-48 I I 9 10 77"4 120-11 much 5-43 65-0 41-12 — — — 40 Researches on Cellulose, II Table III. (Collodion Nitrates). Nitrating mixture Product Propor- ^ a-;? ^^ tion of No. CO »C m g ■^1 -3 HNO.t in O^ 0- 0-; M*: SSj 2 -J acids to r^ Kd Sd is IF ?d 2 |3 1 cellulose P. "S > I 3 62-i8 2i-gi 15-91 13-21 3-20 174 174 30 2 3 61-53 20-02 18-45 12-42 98-70 160 167 30 3 3 60-30 19-71 19-99 11-72 99-28 157 160 30 4 3 58-88 ig-60 21-52 10-96 99-50 148 — 30 5 3 59-77 20-94 19-29 11-74 99-98 159 — 12 6 3 58-34 20-62 21-04 lo-go 99-20 149 — 12 7 3'8 63-84 16-96 ig-20 12-08 — 163 — 30 8 3-8 (i2.-yi 16-46 21-02 11-23 — 153 — 30 9 3-8 63-84 i6-g6 ig-20 11-76 — 156 — 12 10 3-8 62-52 16-46 21-02 10-99 — 151 — 12 II 5 67-60 13-66 18-74 12-43 — 167 167 30 12 5 66-37 13-04 20-59 11-66 — 148 160 30 13 5 64-85 14-90 20-25 10-53 — 140 112 8 14 5 64-11 13-62 22-27 9-35 — — 8 The latter affords an interesting comparison with a pre- ceding series, also directed to the production of collodion cottons, in which the proportion HjSO^ : HNO3 was main- tained at approximately 1:1. The duration of the nitration was 24 hours, at the ordinary temperature. Experiments on the influence of this factor showed that at 35° exactly the same results were attainable in 2 hours. It is to be specially noted that the influence of water is the same, in proportionate effect, as in the case of the mixtures of HjSO^ and HNO3 at I : I. A general account is then given of a series of nitrations in which the starting-point was HNO3 of 1-4 sp.gr. The action of this acid is thus described : ' It causes a partial nitration, but about 63 p.ct. of the cellulose remained un- changed, no oxycellulose being formed here. The product as a whole contained 1-49 p.ct. nitrogen, and the nitrated Synthetical Derivatives — Hydroxy I Reactions: Esters 41 product alone by calculation 4-00 p.ct., which corresponds to dinitrocellulose (Cj^). As this product has never been obtained elsewhere, many attempts were made to isolate it by the action of solvents, or to prepare it otherwise in a purer state, but in vain. By addition of 5 p.ct. H2SO4 some oxycellulose was formed. The product contained about 58 p.ct. unchanged cellulose and 2-15 p.ct. nitrogen; trinitrocellulose (CjJ would contain 5-36 p.ct. nitrogen. Here as well it was not found possible to isolate the products as a chemical individual.' In the following series the conditions were reached in which no unchanged cellulose remained in the products. The HjSO^ was varied, the proportion of HNO3 and HgO being maintained constant at 25 and 24 parts respectively to i part cellulose : Acid mixture Product Yield, p.ct. No. H2SO4 HNO3 H2O Nitrogen, Solubility, p.ct. p.ct. p.ct. p.ct. p.ct. IS 41-86 35-82 22-32 g-27 14-22 134 16 38-47 40-19 21-34 10-32 92-30 142 17 40-83 38-72 20-45 10-76 98-24 151 18 42*92 37-40 19-68 1 1 "02 9890 153 19 48-03 34-18 17-79 12-23 99-58 15s 20 49-37 33-38 17-25 12-77 99-82 166 21 50-71 32-50 1679 13-02 99-32 165 22 52-81 31-27 15-92 I3-II 7-65 167 23 54-92 3006 15-02 i3'45 2-63 173 'The product of No. 15 was treated by a solvent indicated to the author by Bronnert as acting upon nitro celluloses below octonitrocellulose — viz. a solution of 4 parts calcium chloride in 100 parts alcohol of 95 p.ct. This solvent extracted in the cold about 10 p.ct., on boiling with reflux cooler another 35 p.ct., but the remaining product contained almost exactly as much nitrogen as the original product, viz. 9-32 p.ct' (hexanitrocellulose=9-i7 p.ct. N). 42 Researches on Cellulose, II The product of No. 20 has the exact composition of decanitrocellulose, and was perfectly soluble in ether-alcohol. It is therefore evident that the decanitrocellulose exists in a variety of forms, soluble and insoluble. In reference to this point, it is to be noted that from the product of No. 21 to that of No. 22 there is a fall of solubility from 99-3 to 7 -6 p.ct., determined by a slight variation in the composition of the nitrating mixture. In further discussing this question of the relation of solubility in ' neutral ' solvents to composition, it is pointed out that a series of nitrocelluloses soluble in alcohol only result from the action of pure nitric acid on cellulose. Those obtained by the authors were of the com- position of octonitrocellulose (ii'o to ii'i6 p.ct. N). The products resulting from the mixtures of H2SO4 and HNO, at I : I with 15 to 20 parts water, though insoluble in alcohol of 95 p.ct., are soluble in absolute alcohol. Solubility in the latter, however, rapidly decreases with increase of nitrogen ; and 'the true decanitrocellulose,' which is perfectly soluble in ether-alcohol, yields only 1-3 p.ct. to absolute alcohol. Experiments in varying the proportion of ether to alcohol showed that the solvent action of the mixture persists through wide variations from the usual proportion of 3 : i. In reference to the supposed correlation between nitrogen percentage and solubilities, Kisniemsky has endeavoured to reduce this to mathematical expression (Mem. des poudres, 10, 64) based upon the experiments of Bruley (loc. cit.). The author considers that, in view of the experimental results now recorded, the attempt is a priori hopeless, and he has found that the indications derived from Kisniemsky's formula are quite uncertain and in many cases directly misleading. The paper contains the results of a careful investigation of several subsidiary questions, of which we note : 1. The alleged influence of the lower oxides of nitrogen in Synthetical Derivatives — Hydroxy I Reactions : Esters 43 the acid mixture on the composition and proportion of nitro- celluloses resulting from its action : The experiments in elucida- tion of this point consisted in reproducing the results with known mixtures, to which was added N2O4 in known proportions, from o"i3 to 9-28 per 100 of nitrating acid. The results are summed up in the statement that nitrogen tetroxide in the nitric acid, even far beyond the proportion ever found in com- mercial acids, has no injurious action in the manufacture of gun-cotton, neither in nitrogen percentage, yield, nor stability, as measured by the ordinary test. 2. The formation of ' oxycellulose ' in the nitrating process : To detect and also to arrive at an approximate proportionate- measure of oxycelluloses in the nitrated products, the latter were digested for 1 hour at 100° in a standardised solution of methylene blue, and the amounts of colouring matter fixed per I grm. of nitrocellulose were estimated as follows : Nitrogen in nitrocellulose, Methylene blue iixed by p. ct. I grm.-nitrocellulose o-o {original cellulose) o-ooi2 13.65 O'OOIO i3'2i o'ooog I2'76 o-ooii 12-05 0'002I 10-93 0-0036 8-40 0-0I20 From these results it is concluded that oxycelluloses are formed under the action of the more dilute acid mixtures. 3. The behaviour of different commercial grades of cotton in nitration : This point was considered chiefly in regard to the production of collodion cotton. 'Authentic samples of cottons of the most divergent grades were carefully cleaned mechanically and washed in the same way as in the prepara- tion for the manufacture of gun-cotton.' They were nitrated under uniform conditions with the same acid mixture (H2SO4 44 Researches on Cellulose, II 63'84 p.ct., HNO3 16-96 p.ct., HjO 19-20 p.ct.), with the fol- lowing results : No. Commercial designation Nitrogen, p.ct. Solubility in ether- alcohol, p.cl. Yield, p.ct. I 2 3 4 5 Chemically pure surgical cotton-wool American cotton ' middling fair ' American cotton 'Florida ' Egyptian cotton, white, ' Abassi ' Egyptian cotton, natural yellow 11-76 11-56 11-67 11-69 H-61 100 lOO 100 100 100 159 157 153 155 154 It is concluded from these results that variations in ' commer- cial ' quality are without influence on the behaviour of cotton as a chemical individual in the nitrating process. ******* Upon this communication we make the following com- ments : I. The experimental numbers cover an extensive range, and, apart from considerations drawn from the personal reputa- tion of the author, they present, on the internal evidence of the communication, a degree of cohesion and accuracy, both reasoned and objective, which entitles us to deduce a conclu- sion which otherwise would not be warranted. This conclusion is a momentous one in relation to the general argument in Chapter I., p. 5, and we state it in its simplest terms. The stages of nitration of cellulose are not molecular stages, but represent progressive increments of the esterifying groups in a mass-aggregate, which is the reacting unit. If the experimental data are carefully analysed, it will be seen that the number of coincidences with molecular proportions, on the assumed C24 dimensions of the cellulose molecule, are not more conspicuous than the divergences, to meet which we should now require to at least double these dimensions. We need not consider the numbers in detail, which would be merely to reproduce the substance of the paper, but we may take one series as especially illustrative, viz. that in which the effect of the higher tempera- Synthetical Derivatives — Hydroxy I Reactions: Esters 45 tures — 40°, 60°, 80° — in the nitration process are investigated (p. 37). The series of products obtained under this range of variations, which were seven in number, show a nearly con- stant nitrogen percentage, i3-o7-i3-i2, intermediate between the deca- and endecanitrocelluloses (the exact mean being 13-14 p. ct. N), at the same time that the yield of product falls progressively from the theoretical, with the first of the series, to 50 p.ct. of the theoretical with the last. Every circumstance of the original reactions points to an end-product and, even after the fractionation attending the prolonged action at 80°, there is no change of composition. So far therefore, as experimental numbers can establish the point, it would be necessary to admit a definite constant be- tween the stages of the deca- and endecanitrocelluloses ; in other words, to adopt a molecule of n.C4g dimensions in order to formulate the ester in question without employing fractional ratios. Instances of this order can be multiplied to show that the progress of research in this series of esters is obviously rendering ' the cellulose molecule ' a function of the degree of refinement of experimental observations. The question arises. Is it necessary to retain this variable, if not elusive, quantity ? We have already indicated reasons for taking another view of the reacting unit, and we shall return to the question. For the purposes of a general survey of the series of nitric esters, we suggest the following statistical method of expressing their quantitative relations : we may admit as experimentally established that the limit of nitration is the nitrate with i4"i4 p.ct. N. If we assign to this limit the convenient per- centage limit, viz. 100, and express the lower esters by corre- sponding numbers, i.e. as percentages of this maximum, we shall have a nomenclature free from any objection. Since also — -; — = 7-07, the approximate position of a given product is seen by multiplying the nitrogen percentage by 7. 2. A striking point in the series of experimental numbers is the suddenness of the transitions in the associated pro- perty of solubility (ether-alcohol), which are determined by 46 Researches on Cellulose, II variations, relatively minute, of the percentage of water in the nitrating mixture. Thus in the series of p. 40, by increas- ing the water from 15-9 to 18-45, the solubility jumps from 3-2 to 98"7. In the series of p. 36 there are two transitions noted, the increase of water from I4'S to i6'6, causing the solubility to rise from 22 to 99 p.ct. ; and conversely, as the water increases from 20-26 to 33-50, the solubility falls from 99-8 to I -1 5 p.ct. This was a series of experiments in which all other conditions were maintained ' practically constant, and only the proportion of water was varied.' The evidence of these transitions is in favour of a reacting unit varying with the condition of the reaction. There is evidently a critical phase as regards the variant under con- sideration, which is at the proportion of from 15 to 22-5 of water p.ct. on the acid mixture. This phase is also coincident with structural changes in the fibre under nitration. It must be remembered that similar critical phases are met with in many other of the characteristic reactions of cellulose. These phases have their parallel rather in the behaviour of solutions ; and are not consistently interpreted, on the ordinary view of reactions, as taking place between molecular reacting units considered as detached from one another and from the mass- aggregate in which they are associated. If we regard cellulose as a solution-aggregate of units which may be of lesser dimen- sion than ' molecules ' — ions, in fact — the configuration of the aggregate would vary with every change of equilibrium with external conditions. The diminished avidity of the more dilute nitrating acid for the water split ofif in the reaction will permit an increased play of interior hydrolytic changes, and as a result the entire aggregate assumes a different configuration, with no essential variation in the stoichiometrical proportions which govern the reaction. Returning to the experimental facts, it is evident that the author has given positive indications for the study of these phenomena, and the investigation of the products formed at these critical phases must throw a particular light on the funda- Synthetical Derivatives — Hydroxyl Reactions: Esters 47 mental problem. It would be necessary to examine these products so as to fix more exactly their relationships, and to extend the investigation to the celluloses obtained from them by denitration. 3. The author introduces at more than one place the consideration of the lower esters, but the mention of them is only incidental. In view of a theory of cellulose constitution and reactions involved in a general theory of the colloidal state^ it may be shown that the lowest numbers of the series have special significance ; but this scientific interest is over- borne by the fact that they have no ' useful ' — i.e. industrial — applications. The particular methods applied by the author to their examination are not described ; but by implication it appears that, under treatment with the sodium ethylate reagent, nitrates containing 1-5 to 2*15 p.ct. N gave 40 p.ct. of water-soluble products of the saponification — in other words, they are stated to contain ' 60 p.ct. of unchanged cellulose.' If we admit for the moment that these products are mixtures — viz. of di- or trinitrocellulose (C24) with unchanged cellulose — it appears that these lowest terms of the series, which more- over cannot be formulated on the basis of molecular dimensions lower than Cjj, are attacked similarly to the higher nitrate. We note this point as evidence against constitutional formulae of Ce dimensions, such as have been recently proposed (p. 133). But there is an alternative to the author's conclusions which does not appear to have been considered — that is, that the mixtures, if such they be, are in effect homogeneous, representing a solid solution of the esterified groups in the cellulose-aggregate, the properties of which are modified as the properties of a solution would be. This alternative view is adopted, by implication if not in terms, in another publication which we have to notice, and we shall revert hereafter to its consideration. We may add before concluding these obser- vations on the extremely valuable contribution of Professor Lunge, that we have suggested to him, in correspondence, the hypothesis of the ' continuous ' esterification of the cellulose- 48 Researches on Cellulose, II aggregate as an alternative interpretation to that based upon the conception of purely molecular constituent units. The main objection raised by the author is contained in words which we may be allowed to quote : ' It will require very strong evidence to convince me of such a contradiction to all our ordinary chemical notions as a continuous nitration involves. Hitherto I see absolutely no evidence against the natural view that the nitrocelluloses as practically obtained are mixtures of various stages of definite chemical compounds.' It would be necessary to admit the force of this contention so far as regards any experimental mass of cellulose. Such a' mass is made up in the first place of individual fibres, and these are essentially variable. Again, the individual fibres may be resolved into lesser masses, which may be regarded as structural components. The reacting unit we have in mind is of dimensions not necessarily those of such structural units, though definitely related to them ; this conclusion being drawn from the intimate correlation of chemical and structural pro- perties. We have to recognise that a reacting unit may be of variable dimensions andj configuration, as, in fact, the solution of an electrolyte, between the terminals of a current which it conducts, represents some virtually continuous chemical mass in which the ' molecules ' have ceased to exist. As far as we can interpret the consensus of evidence in regard to the colloidal state, it involves some analogous condition of matter, and the properties of colloids exhibiting organised structure mark them out as extreme numbers of a series of infinitely varied and differentiated forms. 2. A LABILE NITRATE OF CELLULOSE. E. Knecht (Berl. Ber. 1904, 549). This communication has an interest which is complementary to the preceding, dealing with the phenomena which attend the formation of the lowest nitrates. Synthetical Derivatives — Hydroxy I Reactions : Esters 49 The experimental observations of the author relate to the ' absorption ' of nitric acid by cellulose from acids of progres- sive concentration, together with the accompanying structural changes in the fibre, as measured by the contraction of the cotton yam which was the form of cellulose employed. The acid 'absorbed,' or rather 'retained,' by the cellulose was measured as such, after allowing the yarn — previously squeezed and pressed to remove the excess of adhering acid — to remain for some days in vacuo over freshly ignited quicklime. The product ' dried ' in this way was resolved by water into cellulose and nitric acid, and the latter was titrated. The following series of numbers resulted : Sp. gr. of acid HNO3 retained by the cellulose Contraction of the yarn, p.ct. I"IOO 3-0 I -200 3-7 i'3oo 3-8 1-325 7-1 1-350 7-4 1-375 7-5 1-380 IO-8 2-5 1-400 27-3 lO-o 1-415 35-8 13-0 An absorption or adsorption compound of the formula CfiHioOs.HNOs gives the weight ratio 100 : 38-8. The author regards a compound of this order, and so formulated, as the initial stage of the true ester reaction. This stage, it will be observed, is confined within a narrow range of variations in concentration of the acid. At the sp.gr. 1-415 the proportion of nitrogen actually combined was found to be 0-5 p.ct. of the cellulose, which was left for 2 minutes in contact with the acid and immediately washed off. When, however, the pro- duct was pressed off and ' dried ' in vacuo under the conditions of the experiments above described, then washed and analysed, II. ^ 50 Researches on Cellulose, 11 it was found to contain 2*2 p.ct. combined nitrogen. This confirms the results of Lunge. It is pointed out that the structural alterations which attend the action of these acids are analogous to those which occur under the action of alkaline hydrates (mercerisation), as indeed was observed by Francis (Chem. Soc. J., 47, 183). Further, that constitutional changes are indicated by the in- creased attraction of the product for atmospheric moisture, which is stated to be 4 p.ct. in excess of the normal for cotton fibre. Lastly, it is suggested that a labile nitrate of this order might be formulated, on the assumption of the cycloketonic formula of Cross and Bevan, as CH2 . (CH0H)4 . Q,(_ ^.Vr "■ ***** «■ * Upon this communication we make the following obser- vations : The author has been impressed with the associated pheno- mena of the reactions, affecting the structure of the fibres, and their general resemblance to those of mercerisation. The contrast is a further illustration of the amphoteric constitu- tion of the cellulose-aggregate. We also note the sensitiveness of the equilibrium of the aggregate in regard to small changes in the concentration of the acid reagents, in which respect also the analogy with the reactions of mercerisation is fully main- tained. If the reaction is approached as a case of equilibrium between two solutions, one of which is the colloidal fibre-sub- stance, a much more promising field for experiment is opened up than by endeavouring to establish fixed points characterised by molecular ratios. No doubt the cellulose undergoes considerable modifica- tion under the reaction, and it would be of interest to establish (i) whether any hydrolysis takes place or whether the hydra- tion changes which obviously occur are merely the result or cause of interior rearrangement, without permanent fixation of Synthetical Derivatives — Hydroxyl Reactions : Esters 5 1 water ; (2) Whether the cellulose, after washing, has reverted to its normal equilibrium or, as is more probable, shows a permanent change of reactivity. The author notes, in effect, the increased absorption of ' hygroscopic ' moisture ; this is probably correlated with increased activity in regard to absorp- tion or adsorption combinations generally ; lastly, there may be more profound changes, such as would modify the reactions of solution in aqueous solvents or of esterification. This series of absorption reactions are obviously continuous with the ester reaction proper. 3 (a). MIXED ESTERS OF CELLULOSE, AND THE INTER- ACTION OF CELLULOSE AND NITRATING ACID. Ueber die Gemischten Ester der Cellulose und das Verhalten der Cellulose zur Niirirsaeure. C. F. Cross, E. J. Bevan, and R. L. Jenks (Berl. Ber. 1901. 34, 2496). This is a continuation of researches on mixed esters, of which certain types were described in the previous volume ('Researches on Cellulose,' 1901, p. 39), notably the aceto- benzoates and nitro-benzoyl nitrates. In view of the well- defined reactions of both sulphuric and nitric acids with cellu- lose, it must be presumed that in presence of a mixture of the acids both will react with the OH groups of the cellulose. Those reactions might be simultaneous and independent, resulting in an equilibrium defined by their relative avidities, or the sulphuric acid might act as ' pioneer,' the SO4H residue being progressively replaced by NO3. The experimental results described in this communication establish the fact of the production of mixed esters. They E2 52 Researches on Cellulose^ II show that in the case of a bleached cotton cellulose the pro- portion of SO4H fixed is highest with the concentrated acid mixture (HjSO^ : HNO3 = 3:1), and diminishes as the mix- ture is diluted with water. In the following series the duration of the nitration was i hour at 16°, and the products were washed for 10 to 15 hours, with successive changes of boiling water, till neutral : Nitrating acid Sulphuric acid in combination as mixed ester. As H:S04 p.ct, of ester Water, p.ct. H2SO4 : HNO3 10' 1 4 2'30 o-o o-o 3 : I 3: I 3: I 3:1 o-oo 0-23 f o"83 I "00 I-3S In a second series the duration of nitration was limited to 7 minutes, and the products were washed with cold distilled water until neutral. Nitrating acid Yield of product Sulphuric acid Water, p.ct. H2SO1 : HNOj Product, p.ct. Original cellu- lose, p.ct. 10-13 2-30 0-0 3 : I 3 : I 3 : I 152-7 l6i-2 134-4 2-56 2-52 4-62 3-91 4-06 6-21 A series of observations on unbleached cotton gave results varying in the contrary direction; highest (4'i7 p.ct. HjSOJ with the diluted mixture, and lower (I'Sg p.ct. HjSOJ with the pure monohydrate mixture. These results were sufficient to establish the fact that sulphuric acid takes part in the ester reaction, in addition to its dehydrating function; which effect has obviously to be taken into account as one of the causes of instability of the Synthetical Derivatives — Hydroxy I Reactions: Esters 53 cellulose nitrates, in the condition in which they are obtained when entirely freed by washing from the uncombined residues of the nitrating acid. 3(^). ON THE FORMATION OF SULPHURIC ESTERS IN THE NITRATION OF CELLULOSE, AND THEIR INFLUENCE ON STABILITY. C. W. Hake and R. Lewis (J. See. Chem. Ind. 1905, 24, 374, and 914). These investigators were led by extensive observations on commercial gun-cottons and derivative products to infer the existence of a cause of instability ' not previously recognised.' They found generally a tendency to liberate sulphuric acid when placed in contact with water, and from the mode of its formation, progressively with successive washings and dryings of a particular gun-cotton, it was inferred that it must result from the hydrolysis of a sulphuric ester. Direct experiments were then made, typical samples of purified cottons being nitrated under ordinary conditions, and afterwards subjected to prolonged washing till free from (soluble) sulphuric acid. Fifteen samples thus prepared, and under varied conditions as regards the nitrating mixture, were analysed and found to contain sulphuric acid (SO^H) in variable proportions, the limiting quantities being 0-22 to 079 p.ct. of the nitrocellulose. The results were confirmed by observations in which Swedish filter-paper was taken as the purest form of cellulose obtain- able, and nitrated under standard conditions. In the fully washed nitrates, the combined HjSO^ was estimated and 54 Researches on Cellulose, 11 found equal to 2 p.ct. of the weight. Various observations were then made to determine the conditions necessary for the elimination of the combined sulphuric acid. It was found to yield very slowly to the action of boiling water. Various acids were tried, but found to have no appreciable solvent action. It was found, however, that when the gun-cottons were slightly moistened (8 to 10 p.ct. water) and exposed to the vapour of acetic acid in a closed space at ordinary temperatures, the ester was attacked and, on subsequently washing with water, sulphuric acid was removed. The action is progressive and several treatments are necessary for the complete elimination of the sulphuric acid. The vapour of nitric acid produces a similar effect, but the action is more rapid ; one treatment for 48 hours effects the complete elimination of the sulphuric residues. The authors consider that their results fully establish the presence of combined SO^H residues in cellulose nitrates prepared with the ordinary acid mixture, and also that the conditions of the ordinary stabilising treatments coincide with those which might be presumed to be necessary for the selective saponification of the sulphuric ester groups. In their second communication {ibid. p. 914) the authors state that, in the discussion which followed the reading of their paper, their attention was called to the earlier paper on the same subject (Cross and Bevan, Berl. Ber. 1901, supra) which, although a careful search was made, was entirely overlooked. Upon these communications we may remark that we have found the study of the nitrocellulose sulphuric esters less promising than that of the aceto-sulphuric esters ; and having in our original communication established the fundamental fact and pointed out its technical consequences, we have prosecuted the investigation by way of the group of analogous Synthetical Derivatives — Hydroxy I Reactions: Esters 55 mixed esters the preparation of which is easily controlled. A full account of these mixed esters is given on pp. 83-92. A study of the saponification of these aceto-sulphates shows that the SO4H residues are relatively resistant to alkaline reagents, from the fact that they combine with the saponifying alkali, unless previously existing in the form of salt (Ca, Mg, Zn). In either case, in saline form these residues are not readily attacked. This justifies the con- clusions of Hake and Lewis, particularly confirmed by G. Macdonald in the discussion on their paper, that a more perfect and rapid stabilisation of nitrocelluloses is secured by keeping up an acid reaction in the waters during the boiling of the esters. The elimination of these residues by exposure to nitric acid vapour is probably the result of esterification by substitution of the monohydric acid. Generally, the substitution of SO4H by NO3 is characteristic of the nitrating process under the usual conditions. Conditions will, no doubt, be found under which this process may be controlled for more exhaustive investigation. It would, of course, be a matter of considerable interest to determine whether these fractional residues of SO^H groups are homogeneously distributed in the sense of modifying the properties of the nitrocellulose as a whole. On the ' mole- cular ' view of cellulose — i.e. of its reacting unit — it would be generally held that these minute proportions merely represent the average composition of a mixture. For reasons previously given, we do not regard this as a necessary interpretation. 56 Researches on Cellulose, II RESEARCHES ON THE STABILITY OF NITROCELLULOSE. Untersuchungen ueber die Stabilitaet von Nitrocellulose. W. Will (Mittheilungen : Neubabelsberg, 2, 1901, and 3, 1902). These important communications are already well known, and have, in fact, procured a general acceptance by specialists of the experimental methods which they describe, as well as of the conclusions which the author draws as to the measure and definition of stability of a given nitrocellulose. It has long been known that, at temperatures considerably below the explosion point, the nitrocelluloses are gradually decomposed with evolution of nitric oxide. To measure the amount of nitrogen thus given off the author devised an apparatus and an experimental method which provides that the nitrogen oxides are carried forward, in a stream of inert gas, over ignited metallic copper, by which they are decom- posed, and the resulting nitrogen (together with any nitrogen possibly formed as an original product of decomposition of the nitrocellulose) measured. The first communication leads to the general conclusion that ' the best definition and test of a stable nitrocellulose is that it should give off at a high temperature equal quantities of nitrogen in equal times.' This is based on the results of an extended experimental investigation in which the test in question is applied to fix the relation of stability to the several essential factors both of the process of nitration and the various methods of after-treatment usually employed to correct the defect of instability which always characterises the original nitrate. The method enables the operator to follow the progress towards the stable con- dition attained under the usual treatment of prolonged boiling, Synthetical Derivatives — Hydroxy I Reactions: Esters 57 or by any alternative process ; this ' stable condition ' becomes, therefore, a ' limiting condition ' (Grenzzustand) or a definite property of the nitrate, and it represents, in fact, its normal instability. For a full account of these investigations we must refer the student to the original paper, or to the abstract in the J. Soc. Chem. Ind. 1901. In the second communication the subject is pursued in reference to the wider and more scientific bearings of the results. In the first section it is more clearly emphasised that the ' limiting equilibrium ' measured in terms of the test is a fundamental property of the nitrate, and the test therefore eliminates all the disturbing factors which may cause variable effects in the methods of testing previously current. In the second section is considered the question of the minimum evolution of nitrogen, which is an invariable for a given nitrocellulose once stabilised, but varies for nitrocellulose produced under varying conditions. Thus, in a series of nitrates produced by acid mixtures in which the ratio H2SO4 : HNO3 was constant — viz. 3 : i — but the water was varied in the percentage proportions 4, 9, 14, the test applied by the several products showed the following quantities of nitrogen evolved in each successive interval of 15 minutes : Nitrogen 0-62 0-56 0-35 Acid mixture used in preparation 4f 9^ I4|- With still further dilution of the nitrating acid and reduction of the sulphuric acid (viz. under conditions of nitration repre- sented by the formula 19-), the N evolution of the stabilised nitrate was further reduced to o-i6. And moreover the stabilisation of this series of gun-cottons by the usual processes of washing and boiling exhibited a graduated stability propor- tional to the reduction of the N-constant. 58 Researches on Cellulose, II It is then shown that the stabilisation is effected by various alternative treatments with water and with alcohol, both at the ordinary boiling temperatures and at 125" to 144°, without affecting the N-constant for a particular product ; further, that when a product is fractionated by suspension in and depo- sition from water, and the fractions successively tested, the N-constant remains unvariable. It has previously been shown that the N-constant — i.e. speed of decomposition — is not definitely related (i) to the actual nitrogen contents of the nitrate, (2) to its solubility in ether-alcohol, or (3) to the physical properties, chiefly viscosity, of its solution. The evidence for the author's definition of the ' Grenz- zustand ' as the most characteristic property of a given nitrate is thus complete. The question of the change from the unstable to the stable condition is discussed in reference to the main cause, which is held to be the elimination of secondary products which act as accelerators of the decomposition under the conditions of the stability test. In addition to the ordinary process of boiling with successive quantities of water, and the special treatments with particular solvents — solvents of the by-products — it is noted in certain cases that, after treating for some time at 135° under the conditions of the stability test and removal of acid products of the decomposition by washing with water, the N-constant falls to the minimum for the product under obser- vation. With certain types of nitrates, on the other hand, the N-constant is unaffected by the treatment. Observations were also made in which the addition of reagents — e.g. KMnO^, KCIO4 — had the effect of reducing the speed of decomposition to the normal minimum. No definite conclusions, however, are stated as to the nature of these secondary influences, nor, more generally, the cause of instability in terms of the normal. Synthetical Derivatives — Hydroxy I Reactions : Esters 59 The probable influence of ' surface ' — i.e. the surface of the nitrates — is incidentally discussed in reference to the action of particular solvents which, whilst undoubtedly removing the by-product causing instability, also affect more or less the surface of the nitrate by the effect of partial gelatinisation. The actual influence of such structural change on the speed of decomposition — i.e. the stability test — ^is referred to further investigation. It is remarked, however, generally that the striking uniformity of the decomposition over a prolonged period of heating is conditioned by the peculiar structure of the cellulose (maintained in the nitrate) which allows large changes in weight, such as attend the decomposition, with no proportional change of surface. The section concludes with a general statement on the technically important question whether stability, as previously defined in terms of the test, should be further Umited to express the condition of the lowest attainable minimum speed of decomposition ? The evidence has shown that in 90 p.ct. of cases investigated the ordinary process of stabilisation leads to a condition of the nitrate in which the ' Grenzzustand ' coincides with this lowest attainable minimum. Of the re- mainder it cannot be said that there is any collateral evidence of non-fulfilment of any of the essential conditions of their employment as explosives. The definition is therefore main- tained as sufficient on all technical grounds. The scientific fact however remains, that the normal instability is expressible in the two ways — i.e. in reference (i) to the uniformity of the decomposition over a lengthened period by heating ; and (2) to the absolute amount of nitrogen lost in the successive unit periods. The speed of decomposition considered in relation to the nitrogen contents of the series of nitrates leads to the general conclusion that with increase of nitrogen up to a nitrate of 6o Researches on Cellulose, II 9 p.ct. nitrogen, the speed of decomposition increases in simple proportion. From 9 to i3'S p.ct. the 'normal in- stability ' increases at a relatively higher rate, and at i3"S p.ct. it is double that which would be calculated from the series of lower nitrates. This relation of divergence has also been found to hold when the temperature of observation in the stability test is raised to 145° and' to 155°. The speed of decomposition as a function of the tempera- ture determining the decomposition leads to some important conclusions. As an example of the relations empirically established, the following tables may be cited : (a) For, a nitrate of i3'i p.ct. N prepared with an acid of formula 9^ : — Temperature Speed of N evolution, mgr. per 15 mins. Quotient for 5" 130 135 140 145 150 155 0-25 0-55 I'lO 2-20 4-15 8-IO 2-2 2'0 2-0 i-g i-g i.e. the speed of decomposition is doubled by each successive 5° of increase of temperature. A further development of this observed uniformity is seen in the following table of results obtained, [b') With nitrates lower in the series : Temperature N percentage of nitrate 973 rss N evolution Quotient for 5° N evolution Quotient for 5° 135 140 145 150 155 160 o'35 068 1-27 2-55 4-90 9-95 I-g I-g 2-0 I-g I-Q 0-13 0-25 0-50 1-05. 2-05 3-95 I-g 2-0 2-1 2-0 I-g Synthetical Derivatives — Hydroxy I Reactions: Esters 6i The identity of the ratio — — — i over an extended series Nt of cellulose nitrates is a further evidence of the invariable characteristics of ' stability ' or the normal instability of these products. It is equally clear that the normal instability, measured in terms of this test, is the correlative of the ulti- mate constitution of these esters. The ratio of N evolution to temperature is expressed by the equation log N = fl + 3. o'9932' in which N = the nitrogen evolved in unit time t = temperature in ° Cent. ; a and b are constants, the values of which have been deter- mined from the results of observations at 145° and 157° Cent. ; a = 8"842ii ■,b = - 22,857 o'9932 is a factor which makes the interpolation formula uniform with that applied by Barboli and Stracciati to express the relations of vapour tensions to temperature (Handb. d. Phys. II. 2, 733, Winkelmann). It is noteworthy that the same formula has been applied to express the dissociation tension of solid ammonium chloride as a function of tempera- ture (Berl. Ber. 2, 139, Horstmann). In the last section the author deals with observations specially carried out to supply data for an equation of the decomposition of the ester. A quantity of lo'o grms. of a given nitrate was heated at 135° in successive periods of 7 hours, until the weight of the residue was constant. The total duration of the heating was 371 hours, the final residue weighing 3-3 grms. and giving the following results of elementary analysis : Calc. for C10H3NO8 C . . . . 44-4 45-3 H . . . . 2-4 I'l N .... 5-3 5-3 62 Researches on Cellulose, II The formula is calculated approximately to permit of a simple expression of the decomposition by the equation : Ci2Hi5(N02)50i„ = CjoHjNOs + 4NO + 6H,0 + 2CO, which is in accordance with direct determinations of the water given off and the ratio of the N evolution to the total loss of weight. The equation, however, cannot be regarded as finally established, and indeed the author's experiments go to show that N and CO2 are simultaneously evolved, and probably also NjO. In commenting upon these important contributions we must avoid as far as possible anticipating the results of future developments on the lines of investigation which the author and his collaborators have initiated. The question of the actual causes of instability of the cellulose nitrates — that is, the particular instability which has long been recognised as merely incidental, and as corrected by the ordinary processes of purification — is, of course, only a side issue of the broader question of the essential or normal instability of those esters. But in view of its technical im- portance and probable influence on the fundamental facts of constitution, it would be a serious omission in future inves- tigations not to take into account the presence of sulphuric residues in the esters as originally prepared. This having now been fully established, together with the fact that the sulphuric acid in the ordinary acid mixture employed plays a direct part in the original reaction, it will be necessary to reconsider the question of stabilisation (i) in reference to the influence of SO4 residues in the ester ; (2) as to whether other incidental by-products contribute ; and (3) whether there may not be some more fundamental factor involved, bound up with the actual constitution or configuration of the ester complex. There is a suggestion of the influence of this third factor in the investigations under discussion. In Synthetical Derivatives — Hydroxy I Reactions: Esters 63 one of the tables (2, p. 10) the minimum N evolution is shown to fall from o'6 to 0-17, as the water in the nitrating mixture employed in preparing the ester increases from 4 to 19 p.ct. But further, whereas the ester prepared with the most dilute acid appears to have been stabilised by merely washing, the others were boiled for 70 to 120 hours in order to arrive at the 'Grenzzustand.' In our own researches, described on p. 52, it was shown that the proportion of sulphuric acid found in the ester reaction was a maximum when an anhydrous acid mixture was employed; on the other hand, with an acid of formula io|-, we obtained an ester which was free from SO4. In the exceptional case noted by Will, therefore, these later results afford an explanation of the direct production of a stable nitrate. But for the low minimum of N evolution in the ' Grenzzustand ' we have to find an explanation. It will be noted that the conditions of nitration in the exceptional case are precisely those in respect of which Lunge says 'at 18 p.ct. water the fibres appear somewhat contracted, and the characteristic twist of cotton fibre is lost.' It is also at or about this point of dilution that sudden changes occur in the physical properties of the nitrates, notably in the solubility in ether-alcohol. It is a coincidence which cannot be overlooked, even though the experimental results recorded by either ob- server are not numerous. Lunge and other observers have fully established the fact that identity of empirical composition, with the associated factor of yield of nitrate, is compatible with considerable divergence in physical properties of these esters ; and to this divergence we have to add the variation in normal instability, independently of the secondary instability due to SO4 residues — i.e. to mixed ester groups, or to by- products of the reaction. This variation is of greater signifi- cance since it is an expression of a more obviously chemical property and of its variations. If we consider the actual reacting units, it would be generally admitted that the typical ester reaction X . OH -t- HNO3 = H^O -1- X . O . NO^ can only take one invariable course. It might be admitted that, in the successive esterification of five OH groups in the Cjj ' unit,' 64 Researches on Cellulose, II there would be a gradation in stability of the combination of the negative groups in the resulting ester; but for the moment we are dealing with the marked variation in stability of types of identical degrees of esterification. The conclusion to be drawn is that the cellulose complex is a variable, under the varying conditions of nitration; that it may assume a configuration in which the OH groups are more basic than in another. Of this, in fact, we have numerous instances in other reactions of cellulose. If this is the true significance of Will's measure of normal instability, it can be verified in more than one independent direction of investigation. Not only can the types showing divergent instability be specially investigated in regard to correlative properties, but other nitrates can be submitted to the dissociation test, more espe- cially a series of metallic nitrates, the elements being selected on the basis of variations of valency and chemical function. But if such variations of normal instability are striking, equally so is the uniformity of the results of the stability test in the majority of cases recorded by Will. Unfortunately, perhaps, the technical objective of the investigation has pre- vented the simultaneous extension of the inquiry to the more fundamental elements of the problems involved — through the investigation of which the actual mechanism of the decom- position would be revealed. The author certainly suggests that the decomposition is primarily one of simple dissociation ; and the attendant oxidations of the carbohydrate nucleus are secondary phenomena. On this view it would be important to measure the factor of surface, and this can easily be accomplished by taking the nitrates in the form of continuous regular threads of definite dimensions. Next, it would be important to ascertain whether any condition is involved comparable with the osmotic pene- tration of ' hygroscopic ' moisture from the surface of cellulose masses to the interior, and vice versa under desiccating condi- tions — whether, to state the point more exactly, the liberation of the NO residues takes place only at the limiting surface, and is preceded by actual migration of the negative groups Synthetical Derivatives — Hydroxy I Reactions: Esters 65 within the mass of the ester. If the decomposition under heat is an exact reversal of the esterification process, we should assume a mechanism of this order. In any case, the experi- mental study of both groups of phenomena leads to the conclusion that they are progressive and continuous, and suggests that we are dealing with a reacting mass of indefinite dimensions rather than with an assemblage of individual molecules. Will has, no doubt with intention, avoided the discussion of these fundamental questions, as the introduction of contro- versial points would tend to obscure the obviously main pur- pose of his contributions, which is to establish an experimental method. He therefore adopts the molecular interpretation of the reactions, and for the sake of simplicity and empirical directness the C12 unit. But the very essence of this new departure is that it introduces a dynamic method of observation, and is a com- plete break with methods of investigation involving only ultimate analytical data and ratios. The same feature we have found to mark the later investi- gations of the esterification process ; and although in this case also the authors endeavour to find a complete expression of their results in interpreting them on the basis of molecular relations, there is evidence to show that these are inadequate and that the phenomena involved are progressive and con- tinuous. The line of progress is thus clearly marked out in attacking cases where the individual peculiarities of cellulose assert them- selves more and more : recognising from the strictly objective standpoint that the only fixed ratios are the stoichiometiical equivalents of the ultimate reacting groups, these reacting groups being associated in a complex of indefinite dimensions and in such a way as to allow of reactions of synthesis and decomposition proceeding progressively and continuously — in which, therefore, molecular dimensions cannot be assigned to the complex or aggregate. II. 66 Researches on ' Cellulose, II 5. SMOKELESS POWDER, NITROCELLULOSE, AND THEORY OF THE CELLULOSE MOLECULE. J. B. Bernadou (London, 1901, Chapman and Hall). This is an essay of some two hundred pages, dealing in the main with problems connected with Service explosives; but from the title it may be gathered that the author travels outside the range of these technical questions to deal with the fundamental problem of the constitution of cellulose and the relation of the nitric esters to the parent 'molecule' or aggregate. The earlier chapters contain the original matter of the work, leading up to the exposition in Chapter IV. of a theory of the constitution of cellulose, of which we shall give a brief account. This is followed by a series of supplementary chapters (Appendix I.-IV., pp. 80-193), which contain the author's translation of well-known papers of Vieille, Bruley, and Mendeljeff on the nitric esters of cellulose, concluding with the reprint of a lecture delivered by himself in 1897 before the United States Naval War College, on ' The Deve- lopment of Smokeless Powder.' The ' theory of the cellulose molecule ' is developed from a generalised view of the specific solvents both of cellulose and of its nitric esters — that is, of the causes underlying these particular solvent actions. The question was approached through a series of original observations on the solubility of nitrocellulose in ether and in alcohol, the results of which are summed up as follows : 1. ' Pyrocellulose ' (the pyrocollodion of Mendeljefif) is readily dissolved in ether upon application of cold. 2. All soluble nitrocelluloses acted similarly in presence of an excess of ether, but some are more readily disintegrated by ether upon application of cold than others. Synthetical Derivatives — Hydroxy I Reactions : Esters 67 3. The necessary degree of cold being developed, soluble nitrocelluloses were not only soluble to any desired extent in ether, but could be colloided directly therein without recourse to liquefaction. 4. The addition of a few drops of alcohol in difficult cases appeared to be equivalent to a lowering of temperature — i.e. in increasing solubility. It appears, therefore, generally that the nitrocelluloses are soluble both in ethyl ether and in alcohol at extremely low temperatures; whereas they readily dissolve at ordinary tem- peratures in mixtures of those liquids. To account for these phenomena the author starts from the general view formulated by Cross and Bevan in reference to the actions of aqueous solvents — that is, that the amphoteric equilibrium of the cellulose-aggregate yields only to the simulta- neous attack of both acid and basic groups in the solvents, the resulting solutions representing a species of colloidal double salt. Both alcohol and ether, as well as cellulose and its deriva- tive esters, are formulated to harmonise with this postulate. We reproduce the author's ' type forms ' : Ethyl alcohol. Ethyl ether. [Under strain in methyl ether.] H H H H CH,-0-CH2 c<; /XC 1 H \0/ H Cellulose. I I HC OH HO CH ! .0^ I J ^H H^ OH HO CH F2 68 Researches on Cellulose, II This latter is termed the unit of polymerisation ; and as the terminology employed is the author's own, we proceed to quote : ' From an examination of the unit of polymerisation it will be seen that the atoms of carbon are united (i) by single bonds with each other, (2) with hydrogen bonds, (3) with hydroxyl bonds, and (4) with what may be termed " water " bonds " of the form (^ \. The free carbon bonds represent polymerisation; the hydroxyl radicles represent nitration in the mean and higher stages ; the water radicles represent mercerisation, and nitration in the lower stage; the open oxygen bond in the "water" radicle represents colloidisation by combination of the half-molecule of the nitrocellulose with the half-molecule of the solvent.' And again : ' If alcohol is capable of effecting the solution of a nitrocellulose, it must effect the solution of both its components or resolve itself into two sub-components, each reacting on one sub-component of the nitrocellulose. These sub-compounds, in the case of alcohol, are / -0\ H,CC and >CH ■^2" The molecule of cellulose should also be represented so as to permit a resolution of the original material into acid and basic sub-components through the disintegration of the water groups. ' These postulates are applied to the experimental pheno- mena of (i) nitration, which is to be regarded as a progressive reaction attended by progressive changes of the reacting mass of cellulose, and (2) hydration. ' Since nitration starts, in accordance with our theory, in Synthetical Derivatives — Hydroxyl Reactions : Esters 69 nitro-substitution in the water radicles ... the hydration should start in the rearrangement of the parts of the molecule adjacent to those water radicles. ... As hydration proceeds, transformation advances towards the total conversion of all the water radicles into hydroxyl radicles. This is accompanied by disintegration of the fibre and a tendency to enter into solution. . . . The molecule splits into two halves, between which there is no chemical union . . . except such as may be represented by electrolytic strain.' We have given a short account of this work so far as it aims to contribute to the main problem presented by the che- mistry of cellulose. The author's views could not be accepted by chemists as a comprehensive statement of the elements of the problem of the constitution of cellulose, nor can we attach any value to the graphic constitutional formulae employed to embody his conclusions, for these resolve themselves as regards cellulose into a hexahydric alcohol CeHg(OH)g, and a view of the fully hydrolysed constituent molecules which is obviously inadmissible. We do not insist that a particular view of a com- plex subject is to be rejected because it is heterodox ; but the heterodoxy must be at least justified by an exhaustive examina- tion of the antecedent orthodoxy. Since this is not in any way attempted, we cannot recommend this section of the author's views to the serious notice of students of the science. But this does not detract from the general value of the contri- bution, which lies in the perspective to which the treatment of the subject is subordinated. Thus ' cellulose ' to the author is not merely a complex chemical individual, but represents the ' organic ' influences under which it was produced. In other words, visible structure or external form are inseparably con- nected with ultimate configuration or chemical structure. The solution phenomena of cellulose and its derivatives are related to constitution in the chemical sense ; also there is a complete external parallelism between the solution of the water-soluble 70 Researches on Cellulose, II forms or derivatives in water and of the esters in their specific solvents, and the causes underlying the interaction between solvent and solute may be regarded as similar. Lastly, the resolved molecules of cellulose and the semi-molecules of the solvents, which are an important feature of the author's hypotheses and arguments, are in effect electrolytic antipodes or ions. These terms, however, are not employed, neither is there any reference in the author's exposition to the general theories of the colloidal state now current. We can only recognise therefore the definite but uncoordinated suggestions of a labile or plastic configuration of the complex, the mental picture of which is quite different from that of an aggregate of hexose molecules or polymeride. As regards the remainder of the work, it is a most useful compilation, especially valuable as a contribution to the tech- nology of explosives. 6(a). ON THE WETTING OF COTTON BY WATER AND BY WATER VAPOUR. O. Masson (Proc. R. S., 1901, 74, 230). The former of these papers (a) embodies the results of a close study of the thermal phenomena attending the absorp- tion of moisture by cellulose. In a preliminary account of previous researches on this subject, the author overlooks the investigations of C. Beadle and O. W. Dahl (Chem. News, ^^896, 73, 180), whose results in terms of gain of weight ■, rise of temperature time of exposure time of exposure expressed in curve form, are in generally close agreement with those of the present paper, and were also extended to various forms of cellulose, including the structureless variety obtained from the sulphocarbonate solution. Synthetical Derivatives — Hydroxy I Reactions : Esters 71 The investigations of Masson have been carried out by means of refined experimental methods, with the purpose of analysing the aggregate phenomena in terms of their essential factors. The thermal changes were measured as a rise of tempera- ture of thermometers carrying the particular mass of cotton under observation wound tightly round the bulb and covering also a portion of the stem. The following remarks are made generally on the observations : Ordinary cotton-wool, and the 'absorbent' modification, behave identically in saturated air ; but when completely immersed in water the latter shows a relatively smaller thermal effect. The rise of temperature reaches a maximum in 5 to 6 minutes on exposure to air saturated with aqueous vapour, 2 to 3 minutes in the case of immersion in water. The author considers that the experiments have proved ' that in both cases the heat-production is due, at all events primarily, to the condensation of water vapour on the surface of the cotton fibres.' On this point it is noted later in the paper, when discussing the experimental results, that the heat evolved per mgrm. of vapour absorbed is 0*57 cal., which is equal to the heat liberated by i mgrm. vapour at 25° condensed to form liquid water at S?""? ; these were the experimental conditions. In this calculation the specific heat of cellulose is taken at o°-336 (Fleury). ' The uncertainties involved in the calcula- tion forbid any further conclusion than that already given, viz. that the heat of absorption by cotton is composed mainly of the heat by liquefaction of water.' The author's further conclusions may also best be given in his own words ; first, in regard to the similarity of the thermal effects under the condition of immersion in water : ' though the absorption of vapour cannot be observed, it must certainly 72 Researches on Cellulose, II occur. . . . The air which adheres to each fibre and fills all interstices maintains that separation of water and cotton sur- faces which is necessary for distillation to occur from one to the other. ' Two questions remain to be discussed, viz. the probable fate of the vapour after it has been condensed in the surface of the cotton, and the bearing of the conclusions here arrived at on the nature of the Pouillet effect in general. ' As to the first question there seems to be three possible views. One of these — the view that it becomes chemically combined to form definite hydrates of cellulose — may be quickly dismissed as there are no facts to support it, and as the evidence points to the absence of any definite limit to the action. ... At the opposite extreme is the view that the con- densed moisture forms and remains as a liquid film on the surface of the solid.' In criticising this possibility the author makes an approxi- mate calculation of the actual surface of a mass, w, of cotton, on the assumption that the constituent fibres are uniform cylinders of diameter i88 x io~^ cm., the area being 4WX i'S2S X i88x io~s= 1,395 ^ cm.D, I -525 being taken as the sp.gr. of cellulose. With w grms. of water deposited, or as in the case of a typical experiment o'liS grm., upon a mass of cotton of 0*900 grm., the thickness of the film would be 15 x io~^ cm. Referring to a paper by G. J. Parks on ' The Thickness of a Liquid Film formed by Condensation at the Surface of a Solid' (PhU. Mag. 1903, p. 517), the above dimensions are considerably outside the superior limit deduced by Parks — viz. 80 X 10"^ cm. for the general case. The author then proceeds to state the third hypothesis ' that the deposited moisture does not all remain as a mechanic- Synthetical Derivatives — Hydroxy I Reactions: Esters 73 ally adherent film on the surface of the cotton, but undergoes continuous osmotic diffusion into the substance of the fibre, and forms with it what may be regarded as a solid solution of cellulose and water.' This, on a general discussion of the experimental results, is found to afford a consistent explana- tion of the phenomena. It may be explained that a considerable portion of this paper is devoted to the statement of the results of the experi- mental observation, their graphic representation by means ot curves, both the absorption effect and the associated thermal effect being so treated ; and lastly, the mathematical analysis of the resulting curves so as to establish the general relation oi the temperature, time, and absorption values. These relation- ships, satisfactorily established as they are inter se, and affording evidence of the value of the author's methods of observation, are not applied to the elucidation of the actual mechanism of the interaction between the cellulose and water, and therefore will not be noticed at greater length. It is only necessary to make it evident that future investigations of the phenomena will be very much facilitated by the author's" contribution both of experimental method and critical treatment of the subject- matter, for the due appreciation of which the original paper must be consulted. 6 {d). INVESTIGATION OF THE HYGROSCOPIC PRO- PERTIES OF NITROCELLULOSE. W. Will (Mittheilungen : Neu Babelsberg, November 1904). Untersuchungen ueber die Hygroskopizitaet von Nitrocellulose. Cellulose. — This study of the cellulose nitrates in relation to 'hygroscopic moisture' necessarily introduces the parent 74 Researches on Cellulose, II substance or original cellulose as a basis of comparison. The investigations herein described differ from those of Masson in being directed mainly to the relations of mass or weight, in limiting conditions of absorption. In a preface dealing with the general question, the author adduces the evidence for regarding this property of attracting moisture from the air as a definitely chemical property of the celluloses. It represents or is associated with a fundamental property of the fibre-substance, or cellulose, independently of its surface and external form. (H. Kuhn, ' Die BaumwoUe, ihre Kultur, Struktur und Verbreitung,' p. 130, A. Hartlebens Verlag, 1892 ; J. Wiesner, 'Die Rohstoffe des Pflanzenreichs,' 2, 179, Engelmann, Leipzig, 1903; Cross and Bevan, 'Cellu- lose,' 1895, p. 4 ; Beadle and Dahl, Chem. News, 71,1; 73, 180 ; C. O. Weber, Ztschr. angew. Chem. 1899, 416.) It has long been generally known that the nitrocelluloses exhibit a similar attraction for atmospheric moisture, though in lower and vari- able degrees, depending mainly upon the degree of nitration of the product ; also that this property and its variations exert an important influence on the ballistic properties of propulsive powders of which these esters are the characteristic constituent. By the investigations of Beadle and Dahl [loc. cit.) it was first definitely shown that in a series of nitrates in which the proportion of ester groups was progressively increased, the proportion of ' normal hygroscopic moisture ' decreased in inverse ratio. The multiplicity of observations to be found in the literature of explosives suggested the necessity of intro- ducing standard methods of observation, and the following are now the normal conditions adopted at Neu Babelsberg : The substance is desiccated in an ordinary drying oven, kept at the constant temperature of 40° ± 1°. It is then exposed to an atmosphere saturated with moisture at 25°, the temperature being maintained at 25° ± 0-5°. At its satura- Synthetical Derivatives— Hydroxyl Reactions: Esters 75 tion-point the substance is weighed in balances contained in cases in which the same conditions of temperature and humidity are maintained. It is obvious that the resulting constants for hygroscopic moisture have no absolute sig- nificance, since both the celluloses and cellulose nitrates are not fully desiccated at 40°. But this temperature is fixed for the reason that the nitrates, and even unstable nitrates, may be safely dried at this temperature without fear of decom- position. Experimental Results. — Of these we reproduce a typical selection : A cotton cellulose dried at 90°, at 70°, and at 40°, other- wise under the same conditions, showed the following relation- ships : Temperature go° 70° 40° Constant weight .... 3'go7 3'924 3*959 P.ct. increase at lower temperatures . — 0'43 1-33 A similar comparison of three raw materials gave the fol- lowing numbers : — Hygroscopic moisture by drying at Temperature of previous drying . American raw cotton Same, converted into high nitrate Same, converted into low nitrate 40° 6-9 1-46 2-38 70° 7-8 I-6l z"6o 100° 8-3 1-68 271 Observations were made upon the rapidity of absorption under the standard conditions above described. The curves obtained for cellulose and for nitrocelluloses confirm those figured by Masson for ' ,, . " ' — , corresponding with a re- a/time latively rapid initial absorption, with an asymptotic approach to the time-axis. The absorption in the first hour exceeds 76 Researches on Cellulose, II one-half the maximum, whereas there is still a measurable increase after 20 hours' exposure. A comparison of a range of raw materials for maximum absorption gave the following numbers : Raw cottons of various origins, all between . 7-3 and 8'o p.ct. Other celluloses and fibious raw materials : Purified cotton rags ... . 6"4 Wood cellulose .... . 6'7-yo Hemp • 77 Jute io'8 Observations were made on various starches, the numbers ranging from io'6 to 11 '4 p.ct., with the exception of potato- starch, which gave i4'i to i6'o p.ct. The influence of various treatments of the celluloses inci- dental to ordinary processes of purification or preparation, gave the following results : ' Beating ' in a paper-making ' Hollander ' increases the absorption from 7-5 to 9-25 p.ct. (averages of four samples) : ether-extraction to a still higher point — viz. 9-7 p.ct. It appears therefore that the normal moisture of a pure cellulose under these conditions is approximately 10 p.ct., and that the moisture complement is influenced by the presence of the residues of oil-wax constituents present in bleached cotton cellulose. Dilute Acids. — Exposure to 2 p.ct. HCl or to 5 p.ct. H2SO4 in the cold reduces the absorption in the ratio ; this is also observed when the treatment with acid rs follows the ether extraction. Dilute Alkalis at boiling temperature produce but slight alteration ; the same holds for the influence of prolonged (50 hours) boiling with water. The following series of determinations show the relative absorptions of various derivative products : Synthetical Derivatives — Hydroxyl Reactions : Esters TJ P.ct. 5-9 i6-i 47 47 Oxycellulose (Nastjukow) .... ' Acid cellulose ' (Bumcke) .... Cellulose acetate (Lederer) .... „ (Weber) .... Lustracelluloses From nitrates denitrated with NH4HS . . . I2'6 From nitrates denitrated with CugClg . . . 11 -8 From highly viscous solution with CujClj . . . 8'g From non-viscous solution with CujClj . . . 15-2 Celluloses from denitration j ^it'o^^IWos^ "'98 P-^'- N 8-2 byCujCIjfrom 1 " "5+ " °;S Cellulose Nitrates. — A selected cotton was nitrated in portions under such conditions that a series of twelve products with N percentages varying from 8*5 to 13 "15 was obtained; for which series the absorptions (i) corresponding to these extremes were determined at 6t6 and i'42 p.ct., and (2) for the intermediate terms a series of intermediate figures such that the sum of the figures for Nitrogen percentage and absorption = i4"6±0"i. The observations were extended to nitrates prepared from a varied range of raw materials, for which this empirical constant was also found to hold good. So also for nitrates obtained by partial denitration of the higher nitrates. The constant was further found to be independent of the factor of solubility in ether-alcohol. Moreover, the rapidity of absorption was found to be uninfluenced by large differences in proportionate solubility. A series of nitrates for which the constant had been deter- mined as normal were dissolved in acetone and reprecipitated by pouring the solution into water. The corresponding structureless modifications retained their water-absorption capacity unaffected by the physical change. Nor was this affected by conversion into /ully gelatinised flakes of i mm.D dimensions. The empirical constant thus fully established for all cellulose 78 Researches on Cellulose, 11 nitrates lying within the range of technical usefulness — i.e. with from 8-5 to 13-9 p.ct. N — obviously fails to hold in the more extreme cases, since in one direction it would imply a nitrate with 14-6 p.ct. N, in the other a cellulose with i4'6 p.ct. hygroscopic moisture. It remains to examine the series of constants in reference to the progressive substitution of the OH groups of the original cellulose (Beadle, loc. cit. ; Cross and Bevan). The progressive substitution of OH by ONOj, in terms of n = number of substituting groups, M = molecular weight of nitrate, N = nitrogen percentage, involves the general expressions N = TOO 324 + 45 « from which, by eliminating n, M = '°°^° , , . . . (i) 3i'ii - N Representing the values for M as a vertical ordinate and the corresponding values of H (hygroscopic moisture) as abscissae, we obtain a straight line expressing the relation M+2S-o7H = S94 . . . (2) and the new relationship H : N takes the form H - 334-3 - Z3'65 N ,. 31-11 - N ■ ■ ' ^^' whence the sum of nitrogen percentage and hygroscopic moisture is no longer constant. For the highest nitrate it has the value i4'i4, reaches a maximum of i4'7i with a nitrate of ii'o p.ct. N, and then falls to 10-78 with «=o, i.e. for the case of the original cellulose. Although such a number is in Synthetical Derivatives — Hydroxyl Reactions : Esters 79 excess of what may be taken as the normal for cotton cellulose, it lies within the range of the author's observed numbers for particular types of cellulose. It may therefore be concluded that the ' hygroscopic moisture ' of the cellulose nitrates is inversely proportional to the number of ester groups. For the nitrates containing from 9-5 to 13-3 p.ct! N the empirical constant, N + 11=14-6 may be taken as representing the experimental facts. Since the celluloses and the nitrates after drying at 40° C. show a further loss (o'3 to 2-0 p.ct.) when fully desiccated over sulphuric acid, it would be necessary in adopting this standard to change the constants in the above expressions (3) and (4) to the following : M + 2071 H = 594; H = 405-8 - 287 N _ 3I-II - N On comparing the absorptions at lower temperatures, viz. at 15° and 5° with those at 25°, which has been adopted as a standard, the following numbers were obtained : — Temperatures N p.ct. of nitrate Moisture absorbed 25° 15° 5° J I3'20 f i"42 27s I 3-00 25° 15° 5° j 12-68 1-95 3-40 350 25° 15° 5° j i2-o6 ( 2-43 4'30 4-61 25° ■? 1 ii-oo f 3-55 5-52 6-01 5° J 9-87 474 7*35 810 8o Researches on Cellulose, II The increased absorption at the lower temperatures does not change the general features of the process. The following are the corresponding empirical constants which obtain at these lower temperatures : At is° . . . H = 2o-i-3N At 5° . . . H = 2i-6-i-4N. It is evident that the relationship between the humidity- equilibrium and temperature is not simple : it is probably complicated by interior hydration, or even hydrolytic changes in the cellulose complex itself. Lastly the author has investigated the equilibrium as effected by varying the proportion of moisture in the atmos- phere to which the nitrates are exposed. The general result of the experimental results is to show that up to 80 p.ct. saturation the ratio of percentage absorp- tion : percentage saturation of atmosphere, is approximately constant. A disproportionate increase in the former occurs as the saturation increases from 80 to 100 p.ct. This section of the subject is still under investigation. We comment on these two communications together for the reason that they are mutually complementary as to subject- matter. It will have been already noted that the authors arrive at opposite conclusions as to the causes underl3dng the common phenomena investigated. We associate ourselves rather with the views contained in Will's communication, con- firming the conclusions which we have previously formulated (' Cellulose,' 1895, p. 4), while very much extending the area of observation. It will further be noted that Masson is chiefly concerned with the thermal phenomena involved, Will with the constants of weight equilibrium. It is superfluous to insist on the importance of the contribution of experimental method and results in either direction of inquiry. As to the mechanism Synthetical Derivatives — Hydroxyl Reactions : Esters 8i of the process, Masson distinguishes between the action at the surface of the fibres and the subsequent process of pene- tration of the interior mass of the fibre substance to form, as he states, a species of ' solid solution ' with or in the cellulose. He also introduces the approximate calculation of the surface of a mass of cotton — incidentally to an argument disposing of the view that the moisture exists as a surface film (see p. 26). With the ' artificial fibres ' at command it should now be possible, by extending the investigations of both authors to these regular forms of the celluloses and cellulose esters, to estimate separately the influence of surface and osmotic penetration. Since these substances can also be produced in film form of any required dimensions, a very wide range of variations of surface to total mass can be controlled, and the investigation thereby much facilitated. Reciprocally, from a more exact knowledge of these factors it may be possible to use the hygrometric observations as a measure of the actual surface of the ' natural ' fibres. Further, a more exact knowledge of the mechanical factors of the absorption will more clearly establish the chemical or molecular features of the limiting equilibrium. While these are generally emphasised by Will, and the final equilibrium is shown to be a property of the substances themselves, inde- pendently of their form or distribution in space, it is evident that this author was unaware of the investigations of Masson and his very opposite conclusions as to the causes, which require a more specific and particular refutation. Outside the question of the causes immediately involved and Masson's summary dismissal of the chemical interpretation of the phenomena, we suggest that a more comprehensive view of these effects will show that they cannot be separated from the general range of hydration-changes determined by alkalis, acids, and salts. The influence of the hygrometric state of the celluloses upon the mechanical properties of their aggregates is an evidence of the intrinsic effects of the combined water ; and it must be remem- bered that the structureless forms of cellulose are more sensitive than are the 'natural' fibrous forms, to these influences. It is II. G 82 Researches on Cellulose, II not superfluous to again insist on our contention that none of these problems can be adequately solved without the inclu- sion of these homogeneous, structureless modifications. The thermal effects noted and measured by Masson are similar in character, and also graduate without break into the effects of contact with dilute hydrolytic agents, acid and alkaline, measured by Vignon ('Cellulose,' 1895, p. 17). These are, no doubt, accompanied by volume changes, graduating into the more extreme effects produced by more concentrated reagents ; and these correlations make a strong a priori case against Masson's interpretations, which present the pheno- mena as quite external to the cellulose. On a priori grounds, also, it is difficult to appreciate the motive for aqueous vapour at a lower temperature steadily migrating to a surface in order to assume the solid state at a higher temperature. It is true that Masson imports into his argument for this interpretation a numerical calculation from experimental data. But these data involve the ' specific heat of cellulose,' for which the doubtful number determined by Fleury (J. Chem. Soc. 1900, 2, 188) is taken. While Masson fully recognises that the factor of the heat of liquefaction of water ' may be supplemented by some other process, either exothermic or endothermic,' which the experi- mental method employed could not test, he makes no sug- gestion that this process could involve constitutional changes in the cellulose substance. In Will's paper, on the other hand, the following passage occurs in discussing the influence of temperature on the hygrometric equilibrium : ' It is not impossible that in the circumstances (of absorption), there enter chemical changes in the molecule of cellulose (anhy- dride formation, hydrolysis of lactonic groups) which may be obscured by the apparently simple relationships between temperature and absorption.' In conclusion, we are of opinion that the interpretations of these phenomena may be unnecessarily obscured by applying to them too exclusively the terminology of molecular com- pounds and the mental pictures which attach thereto. If, on Synthetical Derivatives — Hydroxyl Reactions : Esters 83 the other hand, we regard cellulose as a solution-aggregate in which the molecular forces are merged in new conditions, whether of dissociation or association ; remembering, further, that these conditions have been originally imposed upon the cellulose under the conditions of complete hydration, and, in fact, solution in water ; that the cellulose is both an anhydride by constitution and a dehydrated solution ; that the process of dehydration has taken place under conditions of mechanical strain, which is made evident under ' mercerisation ' and other processes of re-hydration, both by changes of volume and accompanying structural contortions — we have sufficient reason for asserting that the relationships of the celluloses generally to water are the expression of fundamental constitutional facts, which are equally the determining causes of their hygrometric equilibrium with atmospheric moisture. 7(«). CELLULOSE ACETO-SULPHATES. C. F. Cross, E. J. Bevan, and J. F. Briggs (Berl. Ber. 1905, p. 38 (a), p. 1859 (b)). A. In the former paper the general preface connects the re- sults to be described with those of a previous paper, in which the interaction of cellulose with a mixture of sulphuric and nitric acids was proved to result in the formation of mixed esters (Cross, Bevan, and Jenks, Berl. Ber., 34, 2496 ; see also p. 51). Owing to the difficulty of obtaining well-defined products in this series of mixed esters, the investigation was diverted into the more promising region of the aceto-sulphates, which had been- found by preliminary experiments to be extremely well charac- terised derivatives. The method of preparation adopted was a variation of Lederer's process of acetylation, consisting in an increase in the proportion of sulphuric acid. The viscous G2 84 Researches on Cellulose, II solutions which resulted from the reactions were poured into water, the products separating as gelatinous precipitates. These were washed till entirely free from acid. By varying the proportion of sulphuric acid added to the acetylising mixture — a mixture of glacial acetic acid and acetic anhydride in equal weights — it was found that the series of aceto-sul- phates produced grouped themselves on divergent physical characteristics as follows : Low series. — With 5'o p.ct. combined SO4, insoluble in alcohol, soluble in acetone. Middle series. — With 8 to 9 p.ct. combined SO4, soluble in hot dilute alcohol, the solution gelatinising on cooling. Extreme series. — With 25 p.ct. combined SO4, soluble in water, the solutions being highly colloidal. The water-soluble series were only superficially examined ; the lower series were treated by the methods usually adopted in separating and purifying the nitrates, similarly prepared, and were analysed, with results which generally established their relation to the original cellulose — viz. that on a Cg unit three OH groups were replaced. The proportion of SO4 in the middle series corresponded with a formula 4(CeH,0,)SO,(C2H302)jo, which was also in agreement with the yield of product — viz. 174 per 100 of original cellulose. In this preliminary research we observed certain anomalies in the behaviour of the products under treatment for purifi- cation — which, however, did not appear to influence the quantitative relationships of the substituting groups. On further investigation, however, we found that in the course of prolonged washing with water the products underwent changes : in the case of distilled water the esters became more and more Synthetical Derivatives — Hydroxy I Reactions: Esters 85 colloidal, but with an ordinary service-water containing 20/100,000 of CaCOs the change was arrested, and the product resumed its free washing characteristics. These anomalies will be found to receive their explanation by the investigations recorded in the second communication, of which we give a full translation. 7 (V). B. In a recent ' Vorlaufige Mittheilung ' we have described a series of mixed esters resulting from the simulta- neous action of sulphuric acid, acetic anhydride, and acetic acid on cellulose. Owing to the highly colloidal character of these derivatives their analytical investigation involves exceptional difficulty, and in our preliminary researches we became aware of certain variations in their products under preparation — changes in hydration capacity, and probably hydrolysis — which required explanation. The nature of these derivatives and the relationships of the series admit of more precise definition through the results of our continued investigation. In order to avoid experimental difficulties arising from treatment with water, we varied the process of isolating the derivatives by pouring the ' Reaktions-gemisch ' into amyl alcohol, and washing the precipitated product with this alcohol till free from sulphuric acid. The products thus prepared could be dried at 100° C. without change, but they behaved, in contact with water or of alcohol, as free acids, and by quantitative observations the acidity was found to be approximately the equivalent of half the combined sulphuric acid. Hence the inference that, at least in the presence of water, the sulphuric group of these esters takes the form of SO4H. This was confirmed by the preparation of the calcium, magnesium, and zinc compounds by pouring the product of 86 Researches on Cellulose, II reaction into solutions of the acetates of these metals and washing with water till free from soluble impurities. The analysis of these compounds or salts give numbers in satisfactorily equivalent ratios (infra). It appeared therefore that the neutral compounds described in our previous communication were in effect the calcium salts of the acetyl cellulose sulphuric acid now described. These products had been washed with a tap water con- taining 20/100,000 CaCOs, and in the course of a somewhat prolonged washing they gradually took up the equivalent of basic matter. A product so obtained and prepared as a film from alcoholic solution gave the following results on analysis : Total sulphuric acid (H2SO4) . . io'8 p.ct. Total ash 6'5 p.ct., containing CaS04 5-9 „ which represents 40 p.ct. of the equivalent of the total SO4 residue present as S04Ca 2 The following brief statement of the later experimental results, while extending our previous results and modifying our conclusions as explained, will be found to confirm the more important data of our previous paper. Preparation of Acetyl-cellulose Sulphuric Esters AND THEIR SaLTS. In order to obtain more perfect esterification, it was found- to be advantageous to increase the proportion of acetic anhydride to acetic acid from i : i to 2 : i. The result of this variation was an alteration in the reaction equilibrium in the sense that a higher proportion of H2SO4 in the mixture was required to determine the same proportion of H2SO4 combining. Synthetical Derivatives — Hydroxy I Reactions : Esters 87 We may differentiate three well-marked terms of the series of mixed esters which appear to correspond within certain limits to definite stages of equilibrium. (i) Low series (5 to 6 p.ct. combined H2SO4). — Cellulose was treated with 10 times its weight of a mixture of 2 parts of acetic anhydride, with i part glacial acetic acid, containing 4-25 p.ct. of H2SO4. When the solution was perfectly clear the product of the reaction was divided into three portions and poured into {a) Amyl alcohol, (V) A solution of zinc acetate, {c) A solution of magnesium acetate. It was found necessary to employ amyl alcohol for the precipitation and washing of the product (a), since, when washed with distilled water, it cannot be dried without libera- tion of sulphuric acid, which, when heat is employed, mani- fests itself by charring the preparation. The washing with the alcohol was continued for several days, until all traces of sulphuric acid were removed. The products (d) and (c) were Washed with cold distilled water until the washings were neutral. The product (a) of this series, after drying at 100° C, is a white amorphous powder, but in contact with water reacts as an acid ester or acetyl -cellulose sulphuric acid. On boiling with water it is gradually decomposed. It is freely soluble in acetone in the cold and in hot dilute alcohol. The acetone solution yields films of remarkable brilliancy and toughness. The zinc, calcium, and magnesium salts are insoluble in boiling water and dilute alcohol, but they are soluble in acetone. The calcium and magnesium salts are perfectly neutral bodies, and can be dried at 100° C. without change. The zinc 88 Researches on Cellulose, 11 salt, probably on account of the peculiar acid function of the zinc, dissociates on drying even at ordinary temperatures with liberation of free sulphuric acid, and is completely charred when heated at 1 00° C. (2) Intermediate series (the 'normal' products of our first paper) (9 to 10 p.ct. combined HgSO^). — The ester itself and the zinc and magnesium salts were prepared under the conditions above described, but the higher proportion of combined sulphuric acid is conditioned by increasing the pro- portion of sulphuric acid in the acetylating mixture to 6'i p.ct. H2SO4. In this case it is impossible to employ water in isolating the product (a), since it is gradually gelatinised and dissolved by distilled water, even in the cold. The salts (Ca, Mg, Zn) are resistant to cold water, and can be washed to a neutral reaction. The product (a), which in this case also can be dried at ioo°C. without change, is gelatinised by cold water, and is readily Soluble in boiling water, the concentrated solutions becoming gelatinised on cooling ; it is also soluble to some extent in cold strong alcohol, readily soluble in hot dilute alcohol and acetone. The salts are also readily soluble in hot dilute alcohol, and the solutions gelatinise on cooling. The Ca and Mg salts are stable, and correspond with the products described as ' normal ' aceto-sulphates in our previous papers. The Zn salt dissociates on drying, and completely chars at 100° C. High series, water-soluble (25 p.ct. combined HgSO^). — These products were made by treating cellulose with a mixture of equal parts of acetic anhydride and glacial acetic acid con- taining 15 p.ct. of sulphuric acid. The product of reaction was poured into water, in which it was perfectly soluble ; the solution was filtered through cloth to remove residues of cellulose fibres, neutralised with caustic soda, and saturated with pure sodium chloride. Synthetical Derivatives — Hydroxy I Reactions : Esters 89 A precipitate was formed which coagulated when warmed to about 60° C, and which was then filtered off and washed with saturated brine at about 60° C. until the washings were free from sulphates. The product was presumably the sodium salt of an acetyl- cellulose sulphuric acid. It could not be isolated in the pure state, but was pressed and dried, together with the sodium chloride, for the purpose of a preliminary analytical investiga- tion. The percentage of cellulose compound was determined by burning with sulphuric acid, and deducting the NaCl equivalent of the ash from the dry weight. The results of these indirect analyses are given in the subjoined notes (Series 3). This product was completely soluble in cold water, with which it formed a highly viscous, opalescent solution, resem- bling starch paste. It is particularly interesting as showing that we can proceed to high limits of esterification in this series without breaking down the cellulose aggregate, the nearest approximate formula being C6H60(SO,H)(C,H302)3. We are prosecuting the investigation of the series at these high limits so as to obtain derivatives which can be isolated in the pure state for direct analysis. We may again point out that these investigations com- menced with the study of the nitrates (Berl. Ber., 34, 2496), and the proof that, in the ordinary process of preparing the cellulose nitrates with a mixture of nitric and sulphuric acids, the latter acid reacts with the cellulose and SO4 residues are fixed. We also indicated the bearings of this fact upon the important technical question of the stability of these esters. These observations have been fully confirmed and extended 90 Researches on Cellulose, II by the later investigations of Hake and Lewis (J. Soc. Chem. Ind. 1905, 374, 914), see p. 53. We are now sufficiently familiar with the group of aceto- sulphates to investigate mixed esters which contain, in addition, nitric groups. It appears that cellulose nitrates containing 12 to 13 p.ct. N are acted upon by the aceto-sulphuric mixture ; and the resulting product is soluble in dilute alcohol. More- over, the formation of the new product is attended by increase of weight, in spite of the displacement of a portion of the ONO2 residue ; it is probable that these derivatives contain acetyl, sulphuric, and nitric residues. We have therefore a new possibility of bringing the fourth O atom into synthetical reaction. It is clear that until the question of its function has been definitely resolved, we cannot usefully discuss the major question of the actual constitution of the cellulose complex. The production of a tetracetate ^ (Cg) is called in question in recent contributions, and, owing to the exceptional difficulty experienced in the analysis of these esters, the proofs of com- position are lacking in cogency. The solution of this problem is the ultimate aim of these investigations, and we reserve all further discussion until we have extended our investigations of the highest limits of esterification attainable under the conditions of reaction herein described. We subjoin the results of analyses of typical preparations of aceto-sulphates of the three series described in this paper. These preparations, which contained only traces of non- esterified cellulose, were not subjected intentionally to any process of fractionation with a view to separating different components. Their solubilities indicated that they were homogeneous substances, but the possibility of a certain ' A. G. Green in Ztschr. Farb. Textil. Ind., 3, gg, 3og. See also ibii. igg and p. 131. Synthetical Derivatives — H ydroxy I Reactions : Esters 91 amount of fractional dissolution or dissociation during the process of washing is not excluded. Methods. — The hygroscopic moisture was determined by drying in the water-oven for 2 hours at 98° C. Since the zinc salts were charred by this treatment, their dry substance was calculated on the same basis as the magnesium salts. The total H2SO4 was determined after destruction of the cellulose by aqua regia, by precipitation as barium sulphate. The acetic acid was determined by neutralising the products with caustic soda in presence of phenol-phthalein, and then saponi- fying by a known excess of caustic soda in presence of 60 p.ct. (i) Low Series. Product (a) ' ester ' Product (6) zinc salt Product (c) magnesium salt p.ct. p.ct. p.ct. Normal hygroscopic moisture 7-5 decomposed 8-67 Total HaS04 5 "CO 5-96 6-13 Percentage of total HSO re- ) acting as acid to phenol- 40*0 indefinite nil phthalein . Mineral matter (ash) . o-8o (FeA.&c.) } — — Zinc oxide as ZnO 1*93 — Magnesia as MgO — 1-02 Percentage of total HaSO^ saturated by ZnO or MgO — 40-4 42-3 respectively Acetic acid by saponification as CH3COOH . . . ] 52-8 49-8 52-6 Saponified residue. (In this ' series the saponified resi- dues are insoluble in dilute 56'9 59-2 57-6 alcohol, and can be washed free from soluble salts) . j Containing ash sulphates . 4-5 4-30 7-6 Containing combined H2SO4 5-02 3-65 9-57 Percentage of total original H2SO4 remaining in com- bination in the saponified 57-0 35-0 93-0 residue . . . . j 92 Researches on Cellulose, II alcohol. Saponification was allowed to proceed for 40 hours in the cold, and the alkali thus neutralised was calculated in terms of acetic acid, on the assumption that, after the neu- tralisation of any dissociated free sulphuric acid, the residual combined sulphuric acid is very nearly resistant to the saponi- fying action of alcoholic soda. Nevertheless, the values for acetic acid by saponification cannot pretend to any high degree of accuracy ; they indicate however, in an approximate manner, the degree of esterification attained in these com- pounds. (2) Intermediate Series. Product (a) ' ester ' Product (i>) zinc salt Product (c) magnesium salt Normal hygroscopic moisture Total H2SO4 Percentage of total H3SO4 re- ] acting as acid to phenol- phthalein . . .J Mineral matter . Zinc oxide ZnO . Magnesia MgO . Percentage of total H2SO4 ] saturated by ZnO or MgO respectively Acetic acid by saponification 1 as CH3COOH . . . J p.cl. 12-6 lO'oS 50-0 I'OO (Fej03,&c.) 45-0 p.ct. decomposed 8-46 indefinite* 3-36 40-5 51-9 p.ct. 13-2 8-58 nil 1-66 40-6 54-0 * The dry zinc salts contained varying quantities of dissociated free sulphuric acid, in addition to which the ZnO acted as an acid to phenol-phthalein. Owing to the solubility of the saponfiied residue, even in 40 to 50 p.ct. alcohol, these products could not be washed without error, and consequently the combined H3SO4 in them could not be determined. The acetic acid value of the pro- duct (a) is probably too low, owing to the gelatinisation of this product by alcohol, and resulting difficulty of penetration. Synthetical Derivatives — Hydroxyl Reactions : Esters 93 (3) HigA Series. Two different preparations were analysed, giving the following results : Preparation i Preparation 2 Percentage of cellulose compound present in the moist preparation = (dry weight — ■ NaCl calc. from sulphate ash) Normal hygroscopic moisture (calc. on cellulose compound) .... Reaction Combined H2SO4 (calc. on cellulose com- \ pound) Acetic acid by saponification (calc. as CH3COOH on cellulose compound) '1 p.ct. 20-2 Neutral 25-8 49-2 p.ct. 23-9 17-8 Neutral 24-5 29-8 8. ON CELLULOSE-XANTHOGENIC ACID. C. F. Cross and E. J. Bevan (Berl. Ber. 1901, 34, 1513)- Ue^er die Cellulosexanthogensaeure. In a previous publication (ibid. 26, 1090) the synthesis of the cellulose-xanthogenic acid was shown to consist of the two stages : 1. Interaction of cellulose and sodium hydrate in the molecular ratio CgHmOg : aNaOH. 2. Interaction of the alkali-cellulose with carbon disulphide in the ratio CgHjoOg : zNaOH : CSg. The second molecule of NaOH is in direct combination with the cellulose, and the general empirical formula of the xanthate may be written NaS.SC.O.C6H904.NaOH. Direct evidence of the presence of OH groups of, relatively, acid function in 94 Researches on Cellulose, II the cellulose residue, is afforded by the study of the series of those esters representing the several stages of spontaneous decomposition of the original product. In the later stages the ester compounds — i.e. the sodium salts — are insoluble in water, though freely soluble in presence of sodium hydrate. Similar relations are brought out by the author's investiga- tions of the interactions of the alkali cellulose and benzoyl chloride. When the reagents are brought together in the pro- portions QHioOs : 2 -0.2 -5 NaOH : i-o-rs CeHjCO.Cl, the monobenzoate (Cg) is produced without structural modifica- tion of the cellulose fibres. The yield is 8o to 85 p.ct. of the calculated proportion. To determine the maximum of esterifi- cation the reagents were brought together in the proportions prescribed by Skraup (Monatsb. 10, 389), viz. : R{0H)2 : 2[7NaOH : sQHg.COCl], under which conditions the dibenzoate (Cg) is formed, the yield from the fibrous cellulose (cotton) being 90 p.ct. of the theoretical. Its formation is accompanied by structural changes, the dibenzoate being an amorphous product ; and the structural continuity as between fibre and product, which characterises many of the ester reactions of cellulose, is jn this case not maintained. It may be taken as proved that as a maximum proportion two OH groups in the C^ units interact with the alkaline hydrate ; and as regards the further reactions the same general conditions obtain in both directions of synthesis or esterification. Further investigations of the xanthates have shown that the characteristic residue X.CS.S. is of stronger acid function than the monocarboxylic acids of the fatty series. Solutions of the xanthates (viscose) may therefore be acidified with these Synthetical Derivatives — Hydroxy I Reactions : Esters 95 acids — e.g. acetic acid in excess — without decomposing the typical group X.CS.S.Na. Since under these conditions the by-products of the reaction are entirely decomposed and the alkali converted into acetate, the observation has enabled the authors to devise simple analytical processes, by volumetric methods, for estimating the ' xanthate soda ' in such solutions, and thus following the course of the spontaneous decomposi- tions of viscose. This ' reversion ' of the product to cellulose (hydrate) with progressive elimination of the CSS residues, particularly reveals the character of cellulose as a complex, the necessity of studying it as a complex, and therefore of intro- ducing a statistical quantitative control of the analytical deter- minations, which are not required when dealing with simpler ' molecular ' compounds. The following are the analytical methods applicable to this series of compounds : Cellulose may be determined in a weighed quantity of solution (viscose) by (a) ' coagulation ' or precipitation by re- agents, or by heat, (6) spreading and drying down on a glass plate, washing in saturated salt solution, decomposing the xanthate with a mineral acid, washing, purifying, and drying the transparent yJ/zw. Total Alkali is determined by any of the ordinary methods. Xanthate Alkali. — (a) After adding normal acetic acid in quantity equivalent to the total alkali, the xanthate (Na salt) is separated by neutral saturated brine (NaCl), and the uncom- bined acetic acid estimated in an aliquot portion of the filtrate from the xanthate. (6) The xanthate may be titrated with «/io Iodine solu- tion which attacks the xanthate according to the equation : ^„ O.R - ^o/OR . R.0\„„ ^ „ J 96 Researches on Cellulose, 11 The numbers therefore for a pure xanthate are identical with those of titration with an N/io acid (HC1,H2S04). Total Sulphur is determined by oxidising with excess of hypochlorite in alkaline solution, acidifying, filtering, and pre- cipitating as BaSO^. As a result of analytical determinations by the above methods, the following typical characteristics are established : (a) 1 Original product. — The reacting unit is approximately of Cg dimensions, and the xanthate is soluble in, or at least is not precipitated by, neutral dehydrating agents. The aqueous solution, on standing for 24 hours under ordinary conditions of temperature, contains {b) 1 The C12 xanthate, the characteristic constituent of an ordinary ' fresh ' viscose. This body is precipitated by brine or alcohol. (CH— C(OH) )CH2 \CH— CH^ CH— CH/ OH OH OH OH But in view of the exceptionally labile character of the cellu- lose complex, no final value can be assumed for any constitu- tional formula ; the only point to be emphasised is that this lability may extend to the central nucleus of carbon atoms as distinct from migrations of oxygen or hydroxyl groups. Actually, however, this view merges into the wider question which arises on the consideration of the typically colloidal characteristics of cellulose. If these are held to be essentially bound up with the constitution, the nature of this relation- ship becomes a major and previous question. There are particular evidences for the conclusion that molecular con- stitutional formulae cannot represent the condition of the con- stituent groups of cellulose as it exists in the aqueous solution of derivative compounds. These ' solutions of cellulose ' have been previously represented, in accordance with ' molecular ' views, as colloidal ' double salts ' resulting from reciprocal inter- action of acid and basic groups of the ' saline solvent ' and of the cellulose. Much has been added by the investigations, recorded in the present volume to our knowledge of cellulose as typically amphoteric and especially labile in its relation to acid and basic reagents. But we have to apply to these com- plex compounds the general knowledge we have of solutions, and the particular deductions which are now in active circula- tion concerning the colloids as a class. We are thus led to a mental picture of the condition of the constituent groups of cellulose in solution, similar to that which we should form of the solution of a salt of polyvalent radicles, feebly basic and Synthetical Derivatives — Hydroxyl Reactions : Esters 103 acidic. This imports the conceptions of ionisation of reactive constituent groups, which are not the assumed molecules but sub-groups or residues resulting from resolution according to electrolytic actions. From this point of view the state of matter in a ' solution of cellulose ' must necessarily be ex- tremely complex. With dissociation there are the comple- mentary features of association and aggregation, and the functions of water in both groups of phenomena. If these relations are still the subject of active controversy, even where the simplest electrolytes are concerned, we cannot expect to formulate in any detail and for general acceptance a com- prehensive account of the phenomena attending the solution of cellulose. We would make one general statement which may enable us to more clearly define the issue. The crystalline salts may be defined as extrinsic electrolytes — that is, to become conductors of the current they require to change their state, by fusion or solution ; the colloidal salts and amphoteric compounds of the order of cellulose we regard as intrinsic electrolytes. In the separation of a crystalloid from solution the phase of dissociation or ionisation is re- versed, the molecular equilibrium is restored, and the crystal- line form reappears. In the solution of a colloid, in water, for instance, the colloid itself undergoes no essential change ; it is diluted or distended, is homogeneously distributed through the solvent, though perhaps the terms solvent and solute may in such cases be regarded as interchangeable. In the reversal of the process and the separation of the colloid, the solid phase is continuous with the fluid phase : the solid colloid may be said to exist preformed in the solution. The colloid does not consist of electrically neutral molecules, as does the solid crystalline 'electrolyte,' but is formed under the con- dition of electrolytic stress, which stress is retained and ex- pressed in the configuration of the constituent groups. This shows itself under the influence of hydrolytic agents — that is, of H or OH ions — for changes of configuration immediately occur and increase with the duration and intensity of the reaction. We have introduced this general statement in order to I04 Researches on Cellulose, II point out the application of the deductions to the particular case of the cycle of reactions which are under consideration. I. The Alkali-cellulose Reaction presents certain features not brought to light by the investigation of its external aspects, generally summed up in the term ' mercerisation.' Thus it is stated by all investigators of the technical process of ' mer- cerisation ' that the time-factor is of purely subordinate moment ('Die Mercerisation der Baumwolle,' P. Gardner, pp.112-114). But as a matter of, fact the immediate reaction, with its accom- panying structural changes, is succeeded by changes which profoundly modify the xanthate reactions and the properties of the xanthate solution (viscose) obtained. Thus, by merely varying the time-factor, it is possible to obtain viscoses of equal viscosities containing any proportion from 3 to 15 p.ct. of cellulose in solution. So also the progressive advance to a structureless viscose is in direct relation to this time-factor, and the interior changes continue after this phase has been reached, and are then measured by the physical properties (viscosity) and relative stability of the solution. What is the nature of these changes ? In the first place, there is an acidification of the cellulose. The cellulose not only shows a superior acid function in its reactions with bases in presence of water, but manifests an extraordinary resist- ance to esterification. Thus to the acetylating mixture, which depends upon the catalytic action of sulphuric acid (Lederer), the cellulose regenerated from viscose shows itself indifTerent, whereas the original cellulose is rapidly dissolved (as acetate). Even the action of nitric acid may be so regulated as to leave the former cellulose almost unattacked, under conditions which convert the latter into a high nitrate. In the second place, it is to be noted that this change of chemical function is not attended by change of weight of any moment ; it is not a consequence of the fixation or withdrawal of groups reacting with the alkaline hydrate. It is evident that there is a change of configuration both of the complex and the component unit groups, but the nature of this chajige must remain for the present a matter of conjecture. It may Synthetical Derivatives — Hydroxy I Reactions : Esters 105 be pointed out that the acidification of the cellulose is pro- duced by mercerisation under the usual conditions of short exposure to the action of the alkaline hydrate. The progres- sive further modification which follows is only brought into evidence by the xanthate reaction — that is, by the diminishing viscosity of the solutions of the xanthates produced and their increased stability. This suggests a progressive resolution of complex aggregates into simpler component groups ; but we cannot affirm anything as to the nature of this resolution. We might conceive the action rather as a progressive polarisation in which the electro-positive and electro-negative groups take up definitely related positions — the effects of which would be similar to an actual break-up of complexes, such as are held to characterise the hydrolytic resolution of starch. These con- siderations reappear in an analysis of 2. The Reaction of the Alkali-cellulose with Carbon Bisul- phide. — It would appear, for obvious reasons, that the reaction takes place with the electro-positive hydroxyls, with the result of a still further net gain of electro-negative potential and, as a final result, a completely water-soluble saline product. Are we to assume that at any stage there is an actual rupture of these unit antipodes as a result of their varying potentials, or may we assume that they co-exist in polar complexes? Seeing that throughout the whole cycle of reactions of synthesis and reversion the cellulose undergoes no mass or weight changes of any moment ; that all these reactions and changes take place within a circumscribed field — that is, the unit mass of cellulose; that what we have under observation in all our experimental inquiries is to all intents and purposes this unit mass ; that we can only hope to penetrate the present mystery of this complex mass by the application of physical methods, and that up to the present no such method giving direct measurements is available — we may be excused from attempting a final answer to any of these questions. We have raised them for the purpose of exhibiting the complexity of the problems involved, at the same time setting out the lines upon which investigation must proceed. io6 Researches on Cellulose, II 9 (a). INFLUENCE OF CERTAIN REAGENTS ON THE TENSILE STRENGTH AND ON THE DYEING PROPERTIES OF COTTON YARN. J. HUBNER and W. J. Pope (J. Soc. Chetn. Ind. 1903). A necessary basis of this investigation was a preliminary critical inquiry into generally accepted methods of experi- mentally determining breaking-strains, and of reducing the results to a ' mean tensile strength ' with a quantified mean error. The authors finally adopted the method of breaking single threads, in a 'Schopper' machine (Mitt. K. K. Versuchs- anstalt, 1901), taking a large number of observations and quantifying the probable mean error of the mean value. The latter is then reduced to ± 0*7 p.ct. as a maximum. The material to serve as the basis of observation was an Egyptian yarn prepared by boiling with water and sub- sequently with a solution of soda (i p.ct. NagCOj), finally purified by exhaustive washing and air-dried. The comparison of the yarn thus purified with the original raw yarn is given in the following numbers : Tj 1 • ^ • • ^ original 370 -H 17 Breakingstrain m grms. ^ ° 01 — 1 treated 417 + 2'i We now quote the authors in extenso : ' It appears therefore that during the boiling and steeping in water, and the boiling in a i p.ct. sodium carbonate solution, the tensile strength of a raw cotton yarn increases considerably. This increase in strength would seem to be attributable to the boiling in water, and not to the action of the sodium carbonate, which only serves the purpose of removing greasy matters ; this is indicated by a third set of determinations made upon a lea of the material A after it had been scalded. Synthetical Derivatives — Hydroxy I Reactions 107 left for six days in a 15 p.ct. sodium carbonate solution, and washed and dried. 'This gave a mean tensile strength, 4197 ± 2-2 grms., which is numerically identical with that for the purified yarn, 4I7-4 ± 2T grms., and it therefore seems that a 15 p.ct. solu- tion of sodium carbonate has no more action on the tensile strength of the yarn than a i p.ct. solution. ' No satisfactory explanation, based on experimentally de- termined facts, has yet been offered for the considerable changes in the tensile strength of cotton yarn which result from chemical treatment of the yarn ; further, no quantitative explanation of the variation in tensile strength of different threads from the same lea of yarn has yet been given. Now, however, that we have determinations of the tensile strength of yarn possessing a known and high degree of accuracy, we are able to put forward tentatively conclusions concerning the points mentioned, reserving for a later date a more complete experimental study of the conditions than is possible in this preliminary paper. A little consideration will show that the tensile strength of a yarn which contains n fibres in the cross- section is but remotely connected with the tensile strength of the n individual fibres, for, on breaking a thread by the applica- tion of a certain load, the mean effect of the load is to pull the fibres apart, a much smaller number than n being broken, and being broken at different times. So long as the tensile strength of the yarn is less than n times the tensile strength of a single fibre, the strength of the fibres is of quite subordinate import- ance, and the main factor determining the strength of the yarn is the closeness of the twist, and, perhaps also, a specific adhesion or " cling " exerted between the fibres ; it would be expected, therefore, that variations in the twist would show themselves in variations in the tensile strength of the yarn. Mr. W. Myers kindly made a series of 58 determinations of io8 Researches on Cellulose, II the twist on lo-in. lengths of the raw Egyptian yam used in the present work, from which it appears that the twist varies from 22-0 to 30-3 turns per in., the mean being 25"66 turns per in. The twist being the main factor in determining the tensile strength of a cotton yarn, it is interesting to notice that the above three figures relating to the twist of the yarn, viz. 22-0, 30'3, and 25-66, are in the ratio 324 : 446 : 378, which is practically the same as the ratio 330 : 440 : 378, of the lowest, highest, and mean values of the tensile strength determined. This would indicate that for one and the same yarn the tensile strength is directly proportional to the twist. ' On making a count of the fibres in the yarn with which we have worked, it was found to contain an average of 100 fibres in the cross-section. If the above reasoning is correct, each fibre of the raw yarn has a tensile strength greater than 387/100, or 3 '8 7 grms. ; and further, if a series of tests be made under conditions which interpose a resistance to the drawing apart of the fibres, and which increase the number of fibres which have to be broken before the yarn breaks, a great increase in the tensile strength of the yarn should be observed. These conditions would be attained by making the length of thread to be broken less than the average staple of the yarn. We give determinations of the tensile strength of the raw yarn made by diminishing the distance between the clips in the tearing machine to i mm., so that the yarn is broken in lengths of i mm. instead of 100 mm. as in the trials given previously for the raw yarn. ' The tensile strength thus determined on a length of i mm. is 479'4 ± 3'4 grms., whilst that made on a length of 100 mm. is 378-0 ± I "7 grms., so that, as was to be anticipated, the very short length gives a very much higher tensile strength. Determinations of the tensile strength of single fibres, as also Synthetical Derivatives — Hydroxy I Reactions 109 an investigation of the relation between the length of thread broken and the mean tensile strength, are now in progress.' The authors extended their observations to a number of solutions producing effects of ' mercerisation ' more or less pronounced. The results are summarised in the form of tables, which we reproduce, together with the general discus- sion of the results which follow : ' The quantitative results given in the previous pages may be conveniently summarised as in Table A, in which the table from which the results are taken is given in the first, and the tensile strength, t, in the third column. The tensile strength, reduced to that of the standard material, as a basis of 100, is stated in column 4, and the limits of measurement of t and the ratio of those limits, are stated in columns 5 and 6. ' It is interesting to note that the widening of the limits of the observed tensile strengths is accompanied by a com- paratively slight increase in the probable mean squared error of the mean, and that in spite of the wide limits between which the observed tensile strengths lie in many cases, the probable error is never more than 07 p.ct. It is further im- portant to notice that when the mean tensile strength is greatly increased by the application of any chemical treatment which we have studied, considerable disturbances in the structure of the yarn take place, as is indicated by the great widening of the limits of the observed tensile strength in such cases. ' In the previous pages we have not yet discussed the shrinkage which results from the subjection of the standard yarn to treatment ; the lengths of the leas of yarn before and after treatment were ascertained by hanging the lea on a hook, suspending a weight of 25 grms. from it, and measuring the length of the lea as accurately as possible with a millimetre scale. The lengths (I) in centimetres are given in column 4 of Table B. During the determinations of the tensile no Researches on Cellulose, II strength of the various treated yarns, observations were made of the elongation at the breaking load of each thread ; these elongations vary very considerably amongst threads from the same hank, and for our present purpose it suffices to state the mean percentage elongation of the yam at the breaking load. These numbers are stated in column 5 of Table B, and from the lengths of the hanks and the elongations we have calculated the mean lengths (/j) to which the hanks will stretch on application of the breaking load ; a comparison of the values of / and /j shows clearly that after a yarn has been shrunk by mercerisation or other treatment and washed and dried, it cannot be stretched to its original length without breaking. 'Thus 66'o cm. of raw yarn is shrunk to 45 '2 cm. by mercer- ising without tension, washing and drying, and on afterwards stretching the mercerised thread to the breaking-point it only attains a length of 58-0 cm. ; this is a result of peculiar interest in view of the fact that the 66'o cm. thread of raw yarn could be mercerised and washed under tension, and caused to retain that length after drying as mercerised cotton. ' We have indicated above the great probability that the twist is the controlling factor in determining the strength of a yarn, and have given evidence suggesting that for one and the same sample of yarn the tensile strength is directly proportional to the twist. The twist in the samples of yarn produced by the various kinds of treatment mentioned is naturally inversely proportional to the length / of the lea, so that if the law just suggested is also applicable to a yam, independently of the chemical treatment which has been applied, the product of / and / should be constant through- out the series of trials now discussed ; the values oi t % I are given in column 7 of Table B and in column 8 are stated the values of t y. I reduced to the t x /value of the material A Synthetical Derivatives — Hydroxyl Reactions III li coo '^COts.WOO lOISONtN (^VO N OV ts. CO OS »0 txOO ts. CO O OO COOO \0 ONOS 00 lO COts.VO«<0 OivO tN.»O00 tN ts'O VO VO In nS P4 ' H ■^unoxovo t>. in to "I lo to VIVO tsovo ois o,, a= o v^totoo too *OVio tOiOtoO o eS 3- O OCOOO Tt-HvO OOI HiOO CTiO'* Tf »0 lO O* O>00 CO to to (N CSVO OS t-OO CO 2 o b b b b b b b b b b b b b b b b ■;=2 +1+1+1+1+1+1+1+1+1+1+1+1+1+1+1+1 ^s voO^OWHtoos-^ f^oo vo « o H to p •a §-88aaaKs'§22'§a22a ■^ s ts.M OOOOOiO'^eOH OStNOsCOCOLO "m 'm M to CO CO CO « N CO « M CO CO CO CO +I+I+I+1 4-1 +I+I+I +1+1 +I+1+I+I +1+1 fl H ■>~'^- ■ S (M 1 1 "^ -S-l-og -S 5* o .2 ^ " « c-l S r:l * E s S« :ffi . .^■» H?^- -^ ««• iS mS'Z aiU) =mt> 1 P N 0 b b\\b vo H 00 to « M CO O Osoo 00 OS O Ov O OS H H H H X to CO Q 00 tOvO M ■* to tOtoOtoO w O -^OstN. g'S.S w p\ p .■* P Cn t^ p OS « H 00 IN M W H vb CO b t-s In to to ts t>. t^.vo vo vo Is III si ►J " Q'-s,l'f O >>>> "^"^ MXO S -g .">oW - is = 1 112 Researches on Cellulose, 11 {from HI.) as a basis of loo. The values in these two columns show a considerable tendency towards constancy. ' But it is not quite legitimate to expect constancy in the values of Z" X /, because at the breaking-point the load t is not being applied to the length / but to the length /j ; and the product ^ X /j would be more likely to be the constant quantity. This value is given in column 9, and in column 10 is reduced to the / X /j value of the standard yarn as 100; with the exceptions of the hanks 1 treated with potassium or barium mercuric iodides, all the t x l^ values lie between 99' i and 102-3 — 'hat is to say, they are identical within the limits of the experimental error incurred. ' It may thus be safely concluded that although during mercerisation a yarn shrinks by about one-third, its tensile strength remains directly proportional to the twist. 'It should be noted that if, on chemical treatment, the cotton fibre preserves an unaltered volume but merely changes in length, the length / is inversely proportional to the cross- sectional area a of the fibre, so that / x a is equal to the constant volume. But if a given volume of material be made up into rods of different lengths and uniform thicknesses, the breaking loads of the rods will be difectly proportional to the cross-sectional area and inversely proportional to the length. This result is of the same form as that obtained in the present investigation. ' The value of a yarn for manufacturing purposes must be to a considerable extent dependent upon the uniformity of its twist — that is to say, a yarn which will withstand the various operations of bleaching, mercerisation, dyeing, weaving, &c., without giving great trouble owing to frequent breakages, must be one in which the limits of the observed tensile strengths measured on about a hundred separate threads do not lie too far apart. The limits of the observed tensile strength for the Synthetical Derivatives — Hydroxy I Reactions 113 leas of differently treated yam which we have examined are given in Table A, and are conveniently discussed after reducing thepi to the basis of the lowest observed breaking load as 100 ; the quantities thus deduced, which may be termed the ' ratios of the tensile strength limits,' are also given in the table, and will be seen to vary widely for leas which have been subjected to different treatment. Yam No. II., in which the ratio of the limits is as 100 to 133 "3, must obviously give far less trouble during weaving owing to broken threads than that numbered III., in which the ratio is as 100 to i79"o. It seems, therefore, that the manufacturer would derive very reliable information as to the behaviour of a yarn during the subsequent manufacturing processes from the determination of this ratio of the limits of the observed tensile strengths, or by the comparison of the mean tensile strength with the tensile strength observations lying below the mean value. The data obtained in the way which we now describe, treated by the method which we have used, afford specific information as to the strength and the variation in strength of a yarn, whilst the figures given by the lea testing-machine ordinarily used are meaningless in so far as the tensile strength of the yarn is concerned.' 9 {b). THE LUSTRE, THE TINCTORIAL PROPERTIES, AND THE STRUCTURAL ALTERATIONS WHICH RE- SULT FROM TREATING COTTON WITH MER- CERISING AND WITH OTHER LIQUIDS. J. HuBNER and W. J. Pope (J. Soc. Chem. Ind. 1904, 23). This communication aims more directly to elucidate the technology of the mercerising process, but necessarily con- tributes to a knowledge of the primary factors involved in the II. I 1 14 Researches on Cellulose, 11 interaction of cellulose with the alkaline hydrates in presence of water. The main points considered are the specific changes in the properties of the cellulose, resulting from treatment with a series of caustic soda solutions of progres- sively increasing concentration in NaOH — viz. (i) in affinity for direct-dyeing colouring matters ; (2) in dimensions, as measured by shrinkage of a standard cotton yarn ; and (3) in associated structural details. The increase of avidity under (i) was measured by quan- titative dyeings with benzo-purpurin 4 B, using 0-25 p.ct. of the weight of the cellulose. The progressive increase is graphically represented by a curve, generally regular in character, but marked by certain phases of more rapid increase. The experimental observations may be briefly summarised as follows : 1. Solution at 0° to 18° Tw. Effect approximately pro- portional to concentration in NaOH. 2. At 18° to 22°. A more rapid proportional increase of dyeing capacity. 3. At from 22° to 26°, and again from 26° to 30°. Two further stages of increased disproportion. 4. At 30° to 40°. A diminishing effect approximately one- fifth of that in the preceding stages (3), still further diminish- ing at and above 45°. 5. At ss" to 70°. The maximum effect, with very slight variation for varying concentration. 6. At 70° to 80°. There is a reversal of the effect, and cotton treated with the solution at 80° (i*4 sp.gr.) is at about the same level of dyeing capacity as the same cotton treated at 35° (1-17 sp.gr.). (2) The associated structural changes are most obviously demonstrated by shrinkage in length of a standard cotton yarn ; a special method was devised to bring the measure- Synthetical Derivatives — Hydroxyl Reactions 115 ments to a degree of exactitude consistent with the observation of the differences due to increments of 0-005 i" ^^ specific gravities of the solutions of NaOH. The results of observations are given in the subjoined table, which records the reduced lengths of skeins of 200 yards initial length, after treatment for 30 minutes in the cold and measured whilst still moistened with the solution : Strength of soda. Length of hank after Strength of soda. Length of hank after "Tw. treatment °Tw. treatment 198-0 22 171-3 1 196-4 24 163-I 2 195-7 26 160-3 3 195-6 28 i6o-o 4 195-5 30 158-2 5 195-2 35 150-2 6 194-2 40 143-7 7 193-7 45 141-0 8 194-2 50 142-2 9 194-0 55 142-7 10 194-2 60 145-3 12 194-5 65 149-2 14 192-7 70 150-3 16 190-4 75 152-8 18 188-7 80 154-2 20 186-8 The progressive effects may be summarised as follows : A shrinking effect (>i p.ct.) is observed with the weakest soda compared with the action of pure water. From 1° to 20° the shrinkage shows a fairly uniform increase ; at 20° to 22° there is a sudden increase, and again at 2 2" to 24° ; the increase of shrinkage continues at a relatively rapid rate until the maximum is reached at 45°. Above this point of concentra- tion of the NaOH the contraction of the yarn continuously diminishes. The results, plotted as a curve, are seen to be generally coincident with the corresponding variations of dyeing capacity. (3) The changes in minute structure have been very closely 12 1 1 6 Researches on Cellulose, 11 investigated by the authors, and while they confirm the general effects described by previous observers (see ' Researches on Cellulose,' p. 24), they have brought out an additional and special structural feature of the mercerised fibres — viz. the development of spiral but rounded enlargements of the sur- face, which in their opinion are the particular cause of the silky lustre of mercerised cottons. The miscroscopic observa- tions were also extended to the actual process. The action of solutions of varying concentration upon loose cotton-fibres placed under observation in a microscopic field are carefully described. The two main effects observed are the swelling of the fibre-substance and the uncoiling of the fibres, and the authors conclude that for the production of lustre the latter effect must not precede the former. This conclusion was con- firmed by observation of structural changes caused by various reagents in more or less concentrated solution, such as zinc chloride, phosphoric acid, nitric acid (i"4i sp.gr.), sulphuric acid, sodium sulphide, barium, mercuric iodide (saturated solu- tion). The gradation of shrinking and swelling actions of these solutions were examined in reference to gradations of lustre of the treated cottons. Generally, the inferior actions of these reagents, by comparison with those of sodium hydrate, are traceable either to relatively feeble swelling or uncoiling action, or to the fact that untwisting of the f)bre precedes the swelling action. 10. CELLULOSE AND COLOURING MATTER ; THEORY OF DYEING. It is in accordance with the development of our subject -to regard a dyed cellulose as a synthetical compound of cellulose and colouring matter. Even if we gave full value to the con- Synthetical Derivatives — Hydroxy I Reactions WJ tention of controversialists who aim at establishing a mecha- nical or physical ' theory of dyeing,' as against a chemical interpretation of the phenomena, it is obvious that the evidence which leads us to regard the hygrometric equilibrium of cellulose as an essentially chemical phenomenon applies a fortiori to dyeing processes. The combination of cellulose with ' hygroscopic moisture ' is reversible under small variations of temperature : it is in every sense a variable equilibrium. The affinities in play are certainly feebler than in a large number of cases of dyed celluloses, and ' the absence of any definite limit ' is sufficient, according to Masson [loc. cit.), to dispose of the view of the combination resulting from ' che- mical affinity.' While therefore fully admitting the facts which are objected to a chemical interpretation of these residual affinities, we have to ask whether the objectors would maintain their contentions in view of the converging proofs which are accumulating in the investigations of cellulose now under review ? It appears to us that the phenomena which it is attempted to exclusively group under a ' theory of dyeing ' are far too complex and heterogeneous for a satisfactory generalisation. But what is more evident is that this has been attempted be- fore we have come to any agreement as to the fundamental chemistry of the fibre colloids. In the various recent publi- cations on this subject all the conditions of the phenomena are investigated, more or less exhaustively, save and except the all-important factor, the reactivity of the fibre colloid con- sidered generally in relation to the colloids as a class, and, or more particularly, in reference to their many distinguishing characteristics. We were impressed with this want of propor- tion at the time that we wrote upon the subject in 1886, in a series of articles in a technical publication — ' The Dyer ' — and it may be of interest to reproduce the substance of that com- munication. In a preface, the inadequacy of classification of phenomena of this order under the . apparent antithesis ' chemical ' or ' physical ' is set forth, with illustrations drawn from the wide 1 1 8 Researches on Cellulose, II range of hydration effects. The following passage then occurs : ' The study of this one subject of hydration, or water combina- tion, together with the thermal phenomena by which it is attended, leaves us impressed with our inability to draw any line between the several types of combination. . . . ' It is scarcely necessary to point out that the dyeing of fabrics or tissues is only a particular case of the absorption of substances from solution, or, still more generally, of trans- ference by molecular action from one liquid medium to another. ' The simplest case of such a transference is afforded by the removal of a substance dissolved in a liquid by one having a " stronger attraction," into which the substance therefore passes when the two liquids, mutually insoluble, are brought into contact. ' Take the case of the aqueous solution of iodine. This solution, when shaken with carbon disulphide, is deprived of its iodine, which passes over into the latter menstruum, forming a purple solution ; the aqueous solution, on the other hand, being yellow. Should we class these phenomena as physical or chemical ? ' The point thus illustrated by the solution equilibrium of iodine as between water and carbon bisulphide is then further discussed. The particular reactions involved in the mordanting and colouring of the ' organic ' fibrous colloids by compounds of colloidal characteristics, with water as the intervening and con- ditioning agent, can only be investigated on the basis of an understanding of the colloidal state. A general account is therefore given, first of the researches of Graham, and especially of his conclusions as to the continuity of the colloidal state through the soluble and insoluble forms ; secondly, of the researches of van Bemmelen on the withdrawal of acids, bases, and salts from solution by colloidal hydrated oxides (J. Prakt. Chem. 2, 23, 324), with special reference to the quantitative relation of these absorptions to the degree of hydration of the oxides ; thirdly, of the researches of Mills and Takamine on the Synthetical Derivatives — Hydroxy I Reactions 1 1 9 absorption of acids and bases from solution by the • organic ' fibre colloids (J. Chem. Soc. 1883, 153), all of which mani- fest the amphoteric character. In commenting on these results the following passage occurs : ' The authors make no allusion to the change wrought upon the fibres under the action of these reagents, but these, we think, should not be overlooked. We are familiar with the lustre and touch of "scrooped" silk, as compared with the " deadness " in appearance and feel of silk dried from an alkaline solution ; and the effects of acids and alkaline solutions upon cotton, though not so marked, are considerable and of the same order. ' We are not so much concerned with the definition of these changes as "mechanical" or "chemical." Even if slight they are proportionate to the " absorptions " we have been dealing with, and, extending the authors' conclusions, we may class them as chemical. The main conclusion to be drawn is that these fibres or fibre-substances have a certain molecular activity, which manifests itself when they are brought into contact with hydrolytic reagents of the two orders, acid and alkaline.' It is pointed out that these fibrous colloids are capable of an extended series of hydrolyses and that these decompositions are probably preceded by union with the hydrolytic agent, as in the case of the saponification of the fatty ethers, where the emulsion stage probably represents the combination of the complex ester with the alkaline agent, in its hydrated form, preceding the actual resolution into glycerine and fatty acids. ' In conclusion, it is pointed out that the cellulose group supplies a wide range of variations of structural and chemical type, both (i) of ' natural ' origin, and (2) of laboratory produc- tion, the latter including synthetical derivation, and products of resolution by the action of hydrolytic and oxidising agents. Many instances are already known, in which dyeing capa- city varies directly with variations in chemical constitution and independently of structure ; and a ' theory of dyeing ' is I20 Researches on Cellulose, II premature until these complementary phenomena have been more exhaustively studied. We wrote this some twenty years ago and merely reproduce it as a document which, in reference to the controversy still being waged, is an evidence of the slow progress made towards an agreement. We proceed to notice certain more recent contributions to the subject, though not at length, for the reason that the matter must be subordinated to the general perspective of the subject. 10 {a). INORGANIC COLLOIDS AND TEXTILE FIBRES : A CONTRIBUTION TO THE THEORY OF DYEING. W. BiLTZ (Nachr. Wiss., Gottingen, 1904, i). Ueber das Verhalten einiger anorganischen Colloide zur Faser in seinen Beziehungen zur Theorie des Faerbevorgangss. The author has investigated the behaviour of cotton and silk towards colloidal solutions of a widely varied range of inorganic substances, including elements — selenium, tellurium, gold ; oxides — vanadium (pentoxide), molybdenum (' blue '), tungsten (blue) ; and sulphides — cadmium, arsenic, and anti- mony sulphide. These compounds are taken up in varying proportions by the fibre-substances, and their ' adsorption ' was found to have no relation to chemical constitution or function. The dyeing process would appear therefore to be a purely physical effect ; and if the analogy of these inorganic types with the substantive dyestuffs may be assumed, it may be concluded the dyeing properties of these latter are not essen- tially related to their saline constitution. But an experimental inquiry carried out by the Akt. Ges. fiir Anilinfabrikation, Berlin, showed that the inorganic colloids cannot be compared with the substantive dyes in proportionate dyeing effect, neither in Synthetical Derivatives — Hydroxy I Reactions 12 r the quantity • adsorbed ' nor in the resistance of the dyed fibres to washing and to rubbing. It is noted that dyeings with the inorganic colloids can be made much ' faster ' when produced by double decomposition within the fibres. The author draws no conclusion from these observations and leaves the general question entirely open on the alternative issues — i.e. whether colloidal characteristics or saline constitu- tion are the dominant causes of dyeing effects. 10 {b). COAL-TAR COLOURS IN RELATION TO STARCH, SILICA, AND SILICATES. W. SuiDA (Sitzungsber. Wien, July 1904). Ueber das Verhalten von Teerfarbstoffen gegenueber Staerke, Kieselsaeure, und Silikaten. Starch. — A series of dyeings with cold solutions of typical dyestuffs established the following general results : The basic dyestuffs give water-fast colorations ; dyeing power diminishes with the introduction of sulphonic groups. The dyeing effect appears to be referable to a combination of the dye-base with the starch complex ; the acid component is saturated by basic constituents of the ash of the starch, and is left in the dye-bath in the form of neutral salt. The starch behaves similarly to the animal fibres, and to hydro- and oxycellulose, but is differen- tiated from cellulose. Similarly, but conversely, whereas cellulose is dyed directly by the majority of diamine colours, starch is not. The only exception so far disclosed is in the case of the azo-derivatives of diamido-stilbene. Miscellaneous Inorganic Substances, including sulphur, cal- cium, and other sulphates, magnesium and other carbonates, 122 Researches on Cellulose, 11 aluminium oxide and phosphate, zinc oxide, china clay, talc, kieselguhr, pumice, were treated with solutions of basic dye- stuffs, both hot and cold. The silicates alone were dyed. To a typical selection of acid dyestuffs these silicates were inert. Silica and Silicates. — An extended series of experiments are recorded with a careful selection of silicates of known composition. The general result with the basic dyestuffs was the absorp- tion and retention of the base by the acid groups of the mineral, the acid components of the dyestuff being saturated by the basic constituents of the mineral. The dyeings were carried out quantitatively upon a selected china clay, and the proportions of the colour bases fixed were found to be constant. The author concludes finally that 'adsorption' effects play only a subordinate part, and the major effect is a saline combination of dye-base and silica, modified only by the condition of hydration of the mineral. 10 {c). THE DYEING PROCESS IN RELATION TO ACTIVE MOLECULAR GROUPS OF FIBRE-SUBSTANCES. W. SuiDA (Sitzungsber. Wien, January 1905). Ueber den Einfluss der activen Atomgruppen in den Textil- fasern auf das Zustandekommen von Faerbungen. The author made some confessedly unsatisfactory attempts to prepare definite esters of cellulose, and obtained a mono- benzoate and a mono-acetate (Cg) by the action of the anhy- drochlorides in presence of bases ; but the cellulose was simultaneously converted into hydrocellulose. These products, together with a cellulose nitrate (12-1 p.ct. N), were dyed comparatively with a normal cotton in baths of basic, ' direct,' Synthetical Derivatives — Hydroxy I Reactions 123 and azo-acid dyestuffs. No differences were observed in the comparative behaviour of the cellulose and its derivatives. The conclusion drawn is that the OH groups of cellulose are of feeble activity, and that the dyeing of cellulose must be referred to such physical agencies as dissociation, adsorption, capillarity, and solution effects. Silk and wool were treated with acetyl and benzoyl chlorides on the one hand, with alcohols and sulphuric acid and alkyl bromides on the other. The new derivatives were compared with the original fibres for dyeing capacity. The general effect of bringing the acid groups of the animal fibre-substances into previous synthetical reaction, was to suppress their avidity for basic dyes. The dyeing capacity of these products is clearly bound up with ' saline ' constitution and reciprocal interaction with the acid and basic components of colouring matters. 10 (^. THEORY OF DYEING, PART II.; PSEUDO- AND DE-SOLUTION. W. P. Dreaper (J. Sec. Chem. Ind. 1905). This is a continuation of the essay previously appearing in the same journal (1894, pp. 95-101). It is in the main a critical study of recent ' literature on the subject of pseudo- solution, ' surface concentration, ' capillary actions, osmotic phenomena, association and dissociation, generally of the ' physical ' conditions which necessarily accompany dyeing reactions, and which on the author's resume are the effective causes, rather than synthetic reactions, in the chemical sense, between the active component groups of fibre-colloids and colouring matters. 124 Researches on Cellulose, II The final conclusions may be given in the author's own words : ' The conditions of dyeing would seem to be as follows : ' I . A solution of dye or mordant within certain (i.e. definite) limits of aggregation. ' 2. State of fibre corresponding to this degree of aggre- gation. ' 3. Effective " localisation " of the dye, with subsequent concentration of the dye in the fibre-substances. ' 4. De-solution, due to secondary attraction between fibre- substance and dye, or concentration effect. ' 5 . In rare cases primary or chemical attraction may play some part in the process at this stage. ' There is no evidence in favour of solid solution in any of the above results, even in its recent and wider aspect. As is seen dyeing may take place on purely physical lines and, anyway, any chemical action would seem to be of a secondary nature and not essential to the process. ' We have noticed these recent contributions to the subject since they are typical of its literature ; they indicate the present position of the controversy and the conflict of conclusions which prevail. As a general criticism, we may remark that these contributions do not consider, and still less attempt, to exhaust cases which are recognised to be typical, perhaps exceptional. Thus, A. G. Green's primuline dyed upon cotton is a case of special ' affinity ' ; its union with the cellulose is not affected by conversion into the diazo-compound, and this derivative shows a sensitiveness to light very considerably greater than that of the diazo-primuline in the free state. We take this as an evidence of molecular simplicity, of dissociation rather than aggregation. Further, there is no assignable limit to the dilution of a primuline dye-bath which would inhibit its removal by Synthetical Derivatives — Hydroxy I Reactions 125 cellulose in a condition of uniform distribution. In this instance, both as regards the mechanism of the dyeing process and the condition of the colouring matter within the fibre colloid, we recognise a complete inversion of the perspective of principles formulated by Dreaper. The contrast of the dyeing properties of jute and flax enables us to reduce the question of minute structural characters to its true dimensions in relation to operative causes. These fibres are morphologically very similar ; outside the disparity in length and the complexity of the fibre-bundles in jute, there are no marked features of differentiation. Jute, however, has a many-sided avidity for colouring matters. Flax and flax cellulose, on the other hand, are feeble and selective in their dyeing capacities. There is no need to elaborate the purely chemical causes of these divergent relationships. A further contribution to this aspect of the probable causes is the contrast in the dyeing capacities of the ' artificial silks ' and the normal celluloses. If capillary and minute structure played any essential rdle in true dyeing phenomena, the disparity would be in favour of the ' natural ' fibres. But it is in favour of the structureless forms of cellulose (hydrates), and is obviously influenced by two causes which are strictly ' chemical ' — i.e. thei hydration changes which accompany this chemical treatment of synthesis and decomposition, and the change of configuration of the aggregate giving the OH groups a more acid character than the original. As a special type illustrating both the mechanism of the process and the ionising activity of the fibre colloid, we may again cite the interaction of the hgno-celluloses with ferric- ferricyanide. We have seen no reason to change our views as to the essentials of this reaction — i.e. (i) that the ferric- ferricyanide is absorbed as such, as a result of colloidal iso- morphism ; (2) the complex is dissociated by the fibre colloid, owing to the avidity of its ' unsaturated ' quinonoid component for oxygen. Prussian blues are formed in a condition of maximum hydration, and consequently uniformly distributed in solid solution in the fibre-substances. 126 Researches on Cellulose, II Our contention is that the investigation of these and many other cases of this order, upon which there would be, or might be, a preUminary understanding amongst the controversialists as to typical characteristics, would carry us much further than experimental inquiry over a wide range of colouring matters with relatively little attention paid to the special chemistry of the fibre-substance or of the colloidal state. Thus we think the lines of investigation pursued by Pope and Hiibner {loc. cit.), whose contribution requires to be again mentioned in connection with this section of our subject, is much more promising of illustrative results than a diversified research in which no definite account can be given of the variations of the dyed material considered as a chemical individual. It will be evident also, as is in fact generally recognised in all recent contributions to the subject, that the phenomena of dyeing involve as a fundamental factor the rapidly evolving theory of the colloidal state. The most promising phase in this evolution is, in our view, that which we owe to the initiation of E. Jordis, who in the brochure previously men- tioned (p. 6), entitled ' Neue Gesichtspunkte zur Theorie der KoUoide, ' generalises the critical phenomena presented by the colloids as consequences of the modem 'theory of solutions.' A study of this brochure leaves the reader impressed with the fact that the author's exposition implicitly involves the pheno- mena of the dye-bath at every point. As particularly referring to the special points above considered, we reproduce one or two passages : ' Hydrosols are formed and maintained by chemical causes only. In the transitions to hydrogels chemical reactions are mainly concerned, and even in cases where the determining causes appear to be physical, associated chemical factors may be proved to be operative.' ' In consequence of their structure hydrogels manifest an extraordinary develop- ment of (interior) surface, and consequently all those activities which are known as surface actions. ... It cannot be postu- lated that these actions are purely physical in character. Certainly chemical influences are active in the formation of Synthetical Derivatives — Hydroxyl Reactions 127 absorption compounds, and it cannot even be stated that they are inoperative in the surface actions of other substances, such as graphite, charcoal, and fine powders, since Knoblauch has shown that the surfaces of many such bodies are by no means chemically inert.' But in the development of his interesting generalisation Jodis stops short of the fibre colloids — in fact, devotes relatively little attention to the colloidal carbon-compounds. No doubt these alternative provinces will be included in the immediate extension of the theory, and it will come to be recognised that cellulose is a more comprehensive type of colloidal matter than any other known substance. Our own opinion is that the controversial aspect of the ' theory of dyeing ' is the expression of the fact that we find it difficult to avoid applying molecular conceptions and criteria in a region of phenomena where the ' molecule ' never is the reacting unit. We cannot expect to change our perspective by any partial study of new experimental matter. We can only expect to repeat previous experience in finding that conviction will result from the convergence of proofs from various appa- rently independent lines of research. 11. ELECTROLYTIC PHENOMENA. This discussion of the affinities of cellulose and their manifestation in the formation of 'absorption compounds,' leads directly to the consideration of the relationship of cellu- lose to electrical energy. This may conceivably involve a very large question, but for the present we confine ourselves to experimental results, proving that moist cellulose behaves as a conductor of the current, with associated electrolytic pheno- mena of a special character. 11 (a). We gave a preliminary account of such reactions in a paper which appeared simultaneously with our first volume 128 Researches on Cellulose, II on ' Cellulose,' under the title ' Contributions to the Chemistry of Celluloses' (J. Chem. Soc. 1895, 433-451), which we may cite as shortly representing the point of view from which that volume was written. Under the sub-title 'Some Electrolytic Phenomena of Cellulose,' we first described experiments showing that moist cellulose acted as a conductor of current ; secondly, we established the fact that, when a conducting circuit was completed by moist cellulose in contact with two metallic terminals, varying effects were obtained, which could only be explained as resulting from a species of electrolysis of the cellulose, an electrolytic strain under which constituent groups of the cellulose became chemically active. The anode metal was attacked, the reactions with silver being specially marked. Gold and platinum were also attacked. Silver was retained in combination with the cellulose. Gold was in large part carried through the cellulose and deposited in the cathode, platinum to a still greater extent. An instru- ment was devised, described as an ' electrograph,' for demon- strating these effects, and will be found described in ' Nature ' of June 20, 1905. The subject has been more recently investigated by C. R. Darling, and an account of his researches is given in a paper on 'Experiments in Thin-film Electrolysis and a Proposed Application to Printing,' read before the Faraday Society, March 21, 1904. He confirms our results as regards the action on the anode metal, but extends our observations to the important complementary effects which are to be observed at the cathode. Mr. Darling has prepared for us a brief note on these observations, which we give in full : 11 {B). ' When a current of electricity is passed through a sheet of moist cellulose, a latent image of the cathode is retained in the surface layer which' has been in contact with the metal or conductor used as cathode. The presence of the image can be readily detected by brushing the surface of the cellulose with a solution of silver nitrate, and afterwards with hydro- quinone, when a sharply defined picture of the cathode is obtained in silver. Electrographs, or cathode pictures, may Synthetical Derivatives — Hydroxy I Reactions 129 be produced by this treatment in a sheet of paper many months after passing the current; longer contact with the reagents, however, being in this case required. The presence of salts in the cellulose does not prevent the formation of the latent image, which has been detected in a sheet containing a salt of lead after the lapse of five years. Of the possible explanations of the phenomenon, the most obvious, perhaps, would appear to be the presence of matter dissolved from the anode or cathode. This, however, is negatived by the follow- ing observations : (i) the latent image is present when carbon plates are used both as anode and cathode ; (2) there is no tendency for the image to spread on keeping, even when the sheet is moist ; and (3) the image resides in the surface film only, and cannot be detected on the under side of the thinnest sheet. The presence of hydrogen peroxide has been suggested as an explanation ; but not only is it improbable that this substance would be formed at the cathode, but, in addition, a cellulose film containing a latent image may be placed in contact with a sensitised plate for a week without causing photographic action. The phenomenon is therefore entirely different from those investigated by Russell, and referred by him to hydrogen peroxide. The only direct explanation appears to be that the cellulose itself has been modified or strained by the current, so as to render the part in contact with the electrodes more active, chemically, than the remainder ; thus showing, to some extent, the properties of an electrolyte.' These reactions have been investigated rather as ' residual ' phenomena, and in the discussion which followed the reading of Mr. Darling's paper the tendency of criticism was to reject the explanation put forward and substitute those based upon the actions of associated 'impurities.' But in view of the general theory of the colloidal state which has directed our present treatment of the subject, we should regard these phenomena as essentially correlated with the amphoteric activity of the cellulose. Therefore, as the experimental results are clear in their indications, and the alternative hypotheses have been carefully examined and found 130 Researches on Cellulose, II untenable, we are entitled to adopt the conclusions as a con- firmation of the general theory. It will appear that the investigations hitherto carried out have been concerned mainly in establishing the empirical facts ; but in generalising their possible significance and bear- ing on the problems of constitution, a much wider field of experimental investigation is opened up. If the ' cathode image ' is the expression of a permanent stress set up in the cellulose complex, we should be able to ascertain fhe correlative chemical changes. If these should prove to be consistent with variations of constitution or con- figuration such as are determined by the actions of hydrolytic agents, we should have a definite clue to the labile equilibrium of the complex, which has been established by a purely chemical interpretation of these reactions. 11 (c).i Mr. Albert Campbell, of the National Physical Laboratory, has made a number of observations upon the electrical properties of structureless cellulose, as obtained in continuous masses from the decomposition of solutions of the xanthate (viscose). The specimens were prepared by ourselves and represented the approximately ' pure ' cellulose (hydrate). We are indebted to Mr. Campbell, with the kind permission of Dr. R. T. Glazebrook, for the subjoined brief note of results in advance of the fuller publication : ' A number of specimens of pure cellulose in films and sheets of various thicknesses were tested for specific inductive capacity by Maxwell's method with quick periodic charge and discharge, the frequency being varied from 10 to 40 charges per second. In the ordinary air-dry condition the material was found to conduct like an electrolytic solution, showing dis- tinct polarisation. When dried for several hours from 80° C. to 120° C. and tested at temperatures from 120° C. down to 20° C, the apparent capacity fell steadily with the tempera- ^ Report received in course of printing, March 27, igo6. Synthetical Derivatives — Hydroxy I Reactions 131 ture. With long- continued drying the material became very brittle, and in that condition the specific inductive capacity became practically independent of temperature and its value was of the order of 7. These tests would be consistent with the assumption that except when in the dry and brittle con- dition the material may be described by the formula »?(Cellulose hydrates) -I- wHjO where m-\-n = &, constant mass and n increases as the tem- perature rises. This is, in other words, a solution of cellulose hydrates in water, the hydration depending on the temperature, and dissociation equilibrium occurring at any fixed tempera- ture.' 12. ON THE CONSTITUTION OF CELLULOSE. A. G. Green (Ztschr. f. Farb. u. Textil. Chemie, 1904, 3, 97) ; C. F. Cross and E. J. Bevan [ibid. 197) ; A. G. Green {ibid. 309). Ueber die Konstitution der Cellulose. The above contributions comprise (i) a summary of points of evidence as to the constitution of cellulose, (2) a criticism of the constitutional formula proposed in (i), and (3) a rejoinder by the original author. The substance of A. G. Green's review of the evidence may be shortly re-stated as follows : 12 {a). The accepted empirical formula CgHjoOj has given place to one with a C12 molecule, in accordance with the obser- vation of the formation of tri- and penta-nitrates. It is doubtful whether these esters are chemical individuals. The nitrate reaction is complicated by water- exchanges (hydration and condensation), and it is difficult to establish the exact relation- ship of the ester to the original cellulose. K2 132 Researches on Cellulose, II It is assumed that cellulose is a compound of high mole- cular weight mainly on the ground that it is an insoluble colloidal substance. But the reasoning is here insufficient, for alumina and silica, although essentially colloidal, are repre- sented by simple molecular formulae. Failing cogent proof to the contrary, the author adopts the simple formula CgH^jOj. This he proceeds to expand into a constitutional formula to embody the following empirical facts : 1. The highest nitrate has the composition of a trinitro- derivative CgHyOj (O.N02)3. 2. The highest acetate appears to be the triacetate ; the tetracetate, if obtained, is probably derived from a product of hydrolysis. 3. Cellulose combines with sodium hydrate, the product is decomposed by water, the cellulose being converted into a hydrate (mercerisation). This hydrate is more easily soluble than the original cellulose in the aqueous saline solvents, cuprammonium, and zinc chloride. 4. The compound of cellulose and alkaline hydrate reacts with carbon disulphide to form unstable thiocarbonates, from which cellulose is easily regenerated in the form of hydrate. 5. Cellulose does not react with phenyl-hydrazin nor with hydroxylamine. Although therefore not containing CO groups as such, it readily yields products so characterised, under simple hydrolytic treatment. 6. Cellulose is resolved by treatment with sulphuric acid into dextrose as the final product of the attendant hydrolyses. 7. By treatment with hydrobromic acid, cellulose is resolved directly into brom- methyl-furfural (Fenton). 8. By oxidising agents cellulose is converted into oxycellu- lose, a nearly related derivative of weak acidic character, which is decomposed by boiling dilute acids with production . of furfural, and Constitution 133 9. By boiling with calcium hydrate into iso-saccharic acid CH(OH)CH, — COOH I >o CH{0H)CH/ —COOH and dioxybutyric acid (Faber and ToUens). 10. The cellulose nitrates are decomposed by the alkaline hydrates with formation of hydroxypyruvic acid CHjOH.CO.COOH. To summarise these facts the author proposes the constitu- tional formula CHOHC.H. — CH.OH >o >o CHOHC.H/— CHj and discusses its application to the several typical reactions. It is noted in conclusion that the formula can be readily adapted to the case of a more complex constitution — i.e. in the sense of a polymeride. 12 (3). The authors, without impugning the correctness of the deductions contained in the preceding paper, from the experimental data which are stated find it necessary to point out the entire inadequacy of this statement, and generally to show that any attempt to express the chemistry of cellulose in the terminology of molecular formulae is in opposition to the obvious trend of the evolution of research in this province. In regard to the narrower question of a possible formula giving expression to the reactions of cellulose, it is pointed out that the formula proposed by Green is hardly reconcilable with the fact that all the monoses are readily resolved by the alkaline hydrates with production of acid bodies of lower molecular weight, whereas cellulose shows an extraordinary resistance to the alkalis, even surviving ' fusion ' at 180° (Lange). In this respect cellulose is not only far removed 134 Researches on Cellulose, II from the monoses, but from the complex carbohydrates, such as starch, also containing no free CO groups. The Cg reacting unit is further entirely incompatible with the evidences of the series of xanthates and the spontaneous passage of the C,2 xanthate through the stages of C24 to Qj, both of which are well-marked though transitional equilibria (Ber. 1901, 1513). It is further objected that a Cg formula would have at least equal a priori justification in the case of starch. But the quantitative statistics of its various hydrolytic resolutions offer a sufificient evidence of actual complexity. Although there are no directly comparable decompositions of cellulose, those which have been studied reveal a complexity not less pro- nounced. Further, the regeneration of cellulose from the xanthate, traced through the stages of increasing aggregation, is a direct proof of a reacting unit of continuous and large dimensions. It may be objected that the complexity here noted is the expression of physical properties. But the conception of the ' molecule ' — and a simple molecule — is actually derived from physical measurements. It does not therefore seem justifiable to apply these conceptions and terminology in a region where obviously a quite different order of physical relationships obtains. In regard to the number of OH groups in the cellulose unit, and the denial of the existence of a tetracetate (Cg), the authors do not admit that sufficient evidence has been brought forward as against their own quantitative study of the acety- lation — in regard to yield and saponification values of the acetate. But an extension of their investigations to the aceto- sulphates has revealed the production of a well-defined mixed ester of the lowest empirical formula 4C6H702.S04(C2H302)io. in which the ratio 4Ce : SO4 fixes a minimum reacting unit of 4Ce dimensions. 1 ' See also p. 8g and the mixed ester N. [CgHjO . (SO^HXCjHjOa)^. Constitution 135 On the other hand, many of the typical reactions of cellulose may be perfectly well expressed by a Cg formula ; further, it is resolved by certain decompositions to Cg carbohydrates (Flechsig, Stern) or their immediate derivatives (Fenton), and Green's formula is therefore a useful suggestion as to possible constituent groups. But a ketonic constitution is equally if not more probable. In reference to Green's method of disposing of the objection based on the colloidal characteristics of cellulose and its deriva- tives, it is pointed out that the formulae AljOj, SiOj by no means represent the dimensions of the respective oxides. They are really ' valency ' formulae, based upon and derived from the study of other compounds of these elements. The question of the relationship of chemical function to colloidal aggregation is a problem of the future. In cellulose we have extensive evidence that the colloidal properties and structural characteristics of the substance itself and of its derivatives are essentially bound up with chemical constitution. Nothing illustrates this more forcibly than the typical technical methods of obtaining ' artificial silk,' which involve the widely divergent reaction series : Ester Spinning stage Regeneration of cellulose Cellulose Cellulose H2SO4 + HNO. NaOH + CS2 Collodion Xanthate solution Denitration Spontaneous decompo- sition of NH4 salt The close coincidence of structural conditions under these most divergent conditions of chemical modification prove that the colloidal characteristics are not dependent upon incidental ' physical ' factors, but the expression of fundamental proper- ties — i.e. constitution. No subject of experimental inquiry is better calculated to impress the investigator with the essentially chemical factors involved in the colloidal state ; the obvious 136 Researches on Cellulose, II result of recent researches in this special field is to show that cellulose is a prototype of the structural colloids, and investi- gations prosecuted from this comprehensive point of view may be expected to contribute to the progress of science generally. 12 (c). In the rejoinder to the preceding criticism, A. G. Green points out that it is not an essential, nor indeed impiortant, part of his first communication to maintain a Cg formula for cellulose ; his object is rather to formulate the typical Cg unit group. The criticisms, based upon the complex mixed esters and the relations of the xanthate series, do not affect the posi- tion of the constitution of the unit group, and the degree of complexity of the cellulose molecule — i.e. as to how many of the unit groups are united in the complex was left as an open question, with the remark that ' the assumption of a higher molecular weight for cellulose would merely involve oxygen linking of units of the given constitutional formula.' On the special points at issue it is observed that the resist- ance to the action of alkalis is not at variance with the formula, which represents a species of interior aldehyde-ester or acetal. The acetals as a class are much more resistant to alkalis than the aldehydes from which they are derived. On the question of the existence of a tetracetate the author has analysed specimens of the product reputed to have this composition — using the acid method of saponification proposed by A. G. Perkin (Proc. C. S. 1904, 171), and obtained results in close agreement with the theoretical numbers for the tri-acetate. It is also to be observed that the aceto-sulphate formulated in (2) represents the esterification of 12OH group in the 4Cg unit; therefore in the ratio of the tri-acetate. As against a ketone formula it is objected that cellulose does not react with phenyl- hydrazin or hydroxylamine ; and further, that its latent alde- hydic groups are readily brought into evidence under a variety of hydrolytic treatments. Constitution 137 The objection that the formula would equally apply to starch is met by the fact that starch and cellulose show a characteristic difference in their behaviour to hydrobromic acid (Fenton, J. Chem. Soc. 1901, 361), cellulose yielding a large proportion of brom-methyl-furfural, starch but little. The formation of this condensation product was especially taken into account by the author in the formula proposed. We have noted these contributions at some length as they are typical of the two opposite points of view from which the problem of the constitution of the celluloses may be approached. Since the present work is written from the point of view which underlies our criticisms contained in the above article (2), .it would be mere repetition to further discuss the particular point. We make two observations in concluding this notice : 1 . Green's formula may perfectly well apply to constituent groups of cellulose, or may represent those groups in particular phases of decomposition of the complex. The formulation is novel, and of course well grounded ; it is therefore deserving of attention ; but 2 . It is an entirely inadequate representation of the essential characteristics of cellulose as a chemical individual, with asso- ciated physical properties which are inseparable from its con- stitution. This, which is the general basis of our criticism. Green does not refute nor discuss. We may presume that he allows the weight of our contentions on the general question, as we acknowledge the value of his contribution on the parti- cular issue. 138 Researches on Cellulose^ II B. Reactions of Decomposition. From reactions of synthesis, generally involving the reactivity of cellulose as cellulose, we pass to reactions of resolution, in which the cellulose is progressively broken down. 13 (a). THE SO-CALLED ' HYDROCELLULOSE.' A. L. Stern (J. Chem. Soc. 1904, 85, 336). The purpose of this communication is indicated in the opening paragraph, in which the author points out that all kinds of cellulose under the action of certain agents lose their tenacity and become friable. This is usually attributed to their conversion into ' hydrocellulose,' to which Girard assigns the formula C12H22O11. The author's experiments are, however, at variance with the statements of Girard (Compt. Rend., 81, 1105 ; 88, 1322; Ann. Chem. Phys. 1881, 24, 337)- He states that his ' attempts to prepare hydrocellulose by the action of sulphuric acid (i"i45 sp.gr.), and by the action of cold moist hydrochloric acid on air-dried cellulose, proved unsuccessful.' He therefore adopted the plan of boiling with dilute acids — e.g. 5 p.ct. H2SO4. The following effects were established : The conversion into a friable powder was attended by loss of weight, 2 to 4 p.ct. The product again loses weight when the process is repeated. The soluble products of suc- cessive treatments were characterised as 474)- Untersuchungen ueber den Staff und Ettergieumsatz des erwachsenen Rindes. This is a detailed account of elaborate investigations of the total physiological effects of typical fodders, fodder-consti- tuents, and feeding materials generally, the following being more particularly studied : Clover proteids, earth-nut oil, starch, straw cellulose, mo- lasses, meadow hay, oat straw, and wheat straw. The especial interest of the work in its connection with our subject lies in the fact than an artificially prepared straw pulp (cellulose) was included for investigation. This pulp was prepared by treating rye straw under the following conditions: for 1,000 kg. straw, 2,070 litres of a solution containing per litre 55 grms. NaOH, 20 grms. NagCOg, land 22 grms. NagS -f- NajSgOg, treated for 3-J hours at 7 atmospheres steam-pressure, afterwards exhaus- tively washed till free from soluble matter. Such a product JI. L 146 Researches on Cellulose, II may be taken for the purposes of the inquiry as a pure cellulose. It would take us far away from the scope of this work to deal with this important contribution in any detail. We must assume acquaintance with the author's experimental methods and confine ourselves to a brief notice of the final conclusions. These are given in the form of quantitative comparative tables in the concluding section, setting out the values obtained for physiological efficiency (Nutzeffekt) (a) Under the condition of no increase of weight of the animal, (^) in relation to in- crease of weight. (a) In arriving at the ratio of energy utilised in digestion and assimilation to the total available energy of a food, the sources of loss usually taken into account are those of the excreta, solid and liquid, and of the methane and hydrogen resulting from intestinal fermentations. The latter is found, however, to be of small moment and is neglected, as well as the minor thermal changes involved in hydration, hydrolysis, &c., of the food constituents. The following are the numerical results under this head : - Calorific value per 1 grm. Calorific value lost in excreted Total loss Physiological efficiency per I grm. digested Urine Methane Clover proteid . Earth-nut oil . Starch Straw cellulose . Molasses . Meadow hay Oat straw . Wheat straw . 6148 8821 4183 4247 4124 4425 4513 4470 P.ct. 19-3 3-6 8-2 47 5-6 P.ct. lO-I 14-0 12-3 II-5 I2'2 20 -o P.ct. 19-3 lo-i 14-0 15-9 197 i6-9 25-0 4958 8821 3760 3651 3462 35S3 3747 3327 It may be noted that the numbers for the typical carbo- hydrate foods, starch, cellulose, and straw, are nearly identical. Reactions of Decomposition 14; (b) Of the total physiological efficiency the proportion available for increase in body-weight varies between the limits of 17 and 70 p.ct. The following are the percentages deduced as available : Clover proteid Earth-nut oil . Starch . Straw cellulose Molasses Meadow hay . Oat straw Wheat straw . P.ct. 45-2 56-3 58-9 63-1 68-3 40*2 37-6 17-8 And the following table gives the complete view of the utilisa- tion of the energy of the food constituents actually digested : - Losses of energy of digestible constituents Energy equiva- lent of flesh and fat formed In urine Methane fermentation Otlier processes Total Clover proteids . Earth-nut oil . Starch (meal) . Straw cellulose . Molasses . Meadow hay . Oat straw . Wheat straw . P.ct. 19-3 3-6 8-8 47 5-6 P.ct. lO'I 14-0 12-3 9-0 12-2 20-0 P.ct. 44-2 437 37-9 387 26-8 47-0 51-9 6r-2 P.ct. 63-5 437 47-0 457 427 64-8 68-8 86-8 P.ct. 36-5 56-3 53-0 54-3 57-3 35-2 31-2 13-2 On these values the author particularly remarks as follows : The straw cellulose freed from lignone constituents has a nutritive value, as flesh former, equal to that of starch. In comparing the numbers for straw cellulose with those for the untreated straws with special reference to the fate of the furfural-yielding components (furfuroids), it is concluded that the latter are fat formers and are equal in, value to cellulose and starch. It is also concluded that the cellulose is easily digested and in such a way as to prevent the consumption of proteids for respiratory needs, and therefore render them L 2 148 Researches on Cellulose, II available as flesh formers. Lastly, attention is specially directed to the fact that the straws in their raw state have a very much lower efficiency than their component groups; and in par- ticular the separated cellulose, as such, has a much higher actual feeding value than when taken into the organisms in its naturally occurring form. On this communication we make the following remarks : There is a particular interest in the facts revealed with regard to cellulose. According to the ordinary estimate of digestibility, the cellulose might be presumed to be very much inferior in general value to the straw from which it is pre- pared. As the author observes in a more particular discussion of these results, wheat straw with s8'6 p.ct. 'crude fibre' is lowest in the sense of ' efficiency ' — i.e. as a flesh and fat pro- ducer — whereas the cellulose artificially prepared, with 82-1 p.ct. 'crude fibre,' is of efficiency equal to that of a farinaceous meal. In regard to the factor of actual digestibility, on the other hand, these numbers are of no indicative value, for the propor- tions arrived at in the investigations and given on p. 302 of the publication are : — Percentages digested A B Mean Total organic matter Extractive matter 'Crude fibre' . 89-1 77-8 97-0 87-6 807 947 88-3 79-2 95-8 In the case of this ' artificial ' cellulose, therefore, there is not only a high ' efficiency ' or feeding-value, but an unusually high digestibility. In dealing with the causes of differentiation from the original straw, the author mentions the two which are more evident, the altered physical condition by resolution Reactions of Decomposition 149 into ultimate fibres and cells, and the elimination of the relatively indigestible ' lignone and incrusting matters. ' Admitting these factors as contributory, they by no means exhaust the probabilities from the standpoint at which we have arrived. We cannot overlook the constitutional changes which accompany the isolation of the cellulose. We may regard a lignocellulose, as in the straw substance, as a species of ester, and the action of the alkaline mixture used for ' pulp- ing ' the straw as a saponification with secondary actions due to the sulphides used partly of deoxidation, partly of direct synthesis with sulphur groups. But the separated cellulose is not related to the original lignocellulose at all in the same way as, for instance, glycerin to the glycerin ester, from which it is obtained by saponification. As a matter of experi- mental fact, a glycerin separated by an alkaline process is identical with glycerin obtained by the action of acids. But not so cellulose. The cellulose obtained from wood by the action of bisulphites is a different body from ' soda wood pulp ' or cellulose. And further, the ligno- celluloses are not resolved by the purely hydrolysing action of acids. A further condition requires to co-operate with a specific action on the unsaturated ketonic groups of the lignose. Cellulose as a ' solution complex ' or ' intrinsic electrolyte ' must undergo a constitutional modification in response to the action and influence of any and every extrinsic electrolyte with which it is in ' solution contact.' The results of this extensive physiological investigation may both confirm and extend the theoretical views contained in the earlier sections of this treatise. Chemists may be a little ' shy ' of the statistical methods of the physiologist, but this arises from a sense of vagueness or indefiniteness when detached from the familiar conceptions of ultimate molecules and molecular reacting groups. But the statistical method is obviously sound and necessary ; the indis- pensable complement of molecular methods and conceptions in a region where the molecule is indissolubly merged in the aggregates which are the actual objects of investigation. 150 Researches on Cellulose, IT The errors of physiological methods are, no doubt, re- latively large, and there are complications which arise from ignorance and necessary want of control of the vital factors involved. But as every scientific method, and indeed every theory, is justified by its capability of contributing to the advance of exact knowledge, it is evidently gratuitous to criticise physiological methods from any other standpoint. The methods belonging to cellulose chemistry occupy an intermediate position. The statistical method is in this case also the necessary complement of the stoichiometrical, and there is no control of the intermediate phases of reactions, for the reason that there is no knowledge of the actual conditions, material and physical, of the colloids in reaction. It appears therefore that there is much common ground, both of method and material, as between the physiologist and the investigator of the organic colloids. This may be self- evident, but it is none the less true that the region remains to all intents and purposes unexplored. As a practical illustration arising out of the investigations under discussion, if the objective of these investigations were diverted to ' cellulose ' the following problems would arise for investigation : 1 . The relative ' efficiency ' of straw cellulose prepared {a) by the bisulphite process, {b) by the alkaline process. 2. The relative ' efficiency ' of straw cellulose (containing 25 p.ct. furfuroids) and a wood cellulose (containing 6 p.ct. furfuroids). 3. The fate of a structureless cellulose as regenerated from ' viscose ' in the digestive tract of the herbivora, and its quantitative effects in reference to digestibility and 'efficiency.' The results of such investigations should afford an im- portant commentary on the classification of the celluloses previously adopted ('Cellulose,' ed. 1903, 78-87). The diges- tive tract of the herbivora is evidently provided with the condi- tions for both proximate and ultimate resolutions of cellulose. When the proximate resolution is in the ascendant, the cellu- lose appears to have a similar physiological function to that Reactions of Decomposition 151 of starch. This has only been proved in the case of one cellulose, and of that cellulose prepared by a particular process. A comparative investigation of other types would establish some important constitutional relations. In regard to the destructive ferment resolutions with produc- tion of marsh gas, these may be provisionally assumed to be of the type established by the investigations of Omelianski, in which the complementary products are butyric and other fatty acids. A more detailed separation of these factors would be possible, and especially so if the inquiry were extended to other types and groups of herbivorous animals. More definite contributions to the general theory of cellu- lose constitution would result from the study of cellulose derivatives, such as the products of the extreme action of alkalis (mercerisation) and acids ('hydrocellulose'). We have, however, no intention to anticipate these investiga- tions, and the purpose of this superficial forecast is rather to emphasise the intimate connection of our subject with an important branch of physiology, and the certainty of reciprocal advantage to both froni a more definite cultivation of the region common to both. 16. THE DESTRUCTION OF VEGETABLE CELL- WALLS— CELLULOSE FERMENTATION. W. Omelianski (Handb. d. techn. Mykologie (F. Lafar), pp. 245-268). Die Zerseizung der .Baustoffe der Zellwaende der Pflanzen — Die Cellulose- Gaehrung. Tlje author gives a brief historical survey of the earlier observations and investigations of chemists and biologists, E. Mitscherlich (1850), Trecul (1865), van Tieghem (1877)^ Prazmowski (1879). 152 Researches on Cellulose, II More definite results come from the researches of PopofF, Tappeiner (1880), and Hoppe-Seyler (1886), which were chiefly directed to the chemical factors and gave general in- dications of the main products — viz. as gaseous end-products, methane and carbonic anhydrate; and as intermediate pro- ducts, acetic and butyric acids. Parallel with these investigations the subject was attacked by agricultural chemists in reference to the changes taking place in farmyard manure when stored in masses, but beyond establishing the production of methane, and general statistics of the destruction of the straw substance, these researches did not contribute definite results as regards cellulose. It was reserved for the work of Oinelianski to establish a well-defined method of observation leading to exact results. This author's researches were confined to the normal celluloses, employed in the form of Swedish filter-paper. The cellulose ■was placed in flasks which were filled with a nutrient solution containing, per litre, potassium phosphate i grm., magnesium sulphate o'5 grm., ammonium sulphate i grm., and a trace of sodium chloride. The bacteria were introduced in an inoculating liquid obtained from horse-dung or river mud. Chalk was present in excess throughout the period of fermentation. A tem- perature of 34° to 35° was found to be the optimum. A general preliminary investigation established two types of fermentation, the one characterised by the production of hydrogen, the other by the evolution of marsh gas, as the main gaseous product. Smce the incubation period required for the former is much shorter, and the bacilli are sensitive to a temperature of 75", it is easy to inhibit this type and procure the hydrogen fermentation by heating the inoculating fluid for a short time at that temperature. Reactions of Decomposition 153 The following are the quantitative results obtained in typical cases of either fermentation : Hydrogen Fermentation. Cellulose taken „ residual „ fermented grms. 3"4743 0'1272 3'347i Products. Fatty acids Carbonic acid . Hydrogen grms. 2*2402 0*9722 0"0i38 3*2262 The fatty acids consisted of acetic and butyric acids in the molecular ratio 1*7 : 1'o. The fermentation period in this experiment was 13 months. Methane Fermentation. Cellulose taken „ residual „ fermented grms. 2*0815 0*0730 2*0065 Products. Fatty acids Carbonic acid . Methane grms. 1*0223 0*8678 0*1372 2*0273 The fatty acids consisted of acetic and butyric acids in the molecular ratio 9:1. The fermentation period in this experi- ment was 4^ months. From these investigations we are unable to draw any conclusions as to the intermediate products of resolution. We have previously had the opportunity of contributing material for the similar investigations of A. Macfadyen and F. R. Blaxall, the results of which are published under the descriptive title 'Thermophilic Bacilli,' J. Path. Inst. 1894 and Trans. Jenner Inst. (2), p. 162. The material which we supplied was a typical range of fibrous celluloses and also structureless cellulose (hydrate), all of which were attacked and resolved, in periods of from 7 to 2 1 days, under the conditions of these authors' experiments. We were unable to detect any 154 Researches on Cellulose, II intermediate carbohydrate products in the solutions which we examined. These investigations have been resumed, and we have again had the opportunity of co-operating in regard to the supply of experimental material ; also of confirming the obser- vations of Omelianski (loc. cif.) as to the visible evidences of the progressive destruction of the cellulose in the case of the fibrous celluloses. The microscopic observations of this observer are photographically recorded, and they will be found to faithfully represent the structural changes which the celluloses undergo. The only point which calls for particular remark is that the structural changes indicate a localised attack of the organisms, which implies that the breakdown of the cellulose requires actual contact with the bacteria. This interpretation is also consistent with the time-factors of these fermentations, and the general absence of soluble carbohydrate products. It is, however, impossible to affirm that an intermediate stage of solubilisation, as a result of enzyme action, does not occur. Attention is being especially directed to this point in the further investigations now in progress, as well as to the more specific control of the biological features (Macfadyen). Pending the results of these developments, we must accept the inconclusive position as to the actual mechanism of the process. So far as the evidence has accumulated, it is con- sistent with the view that, as a hydrated colloid, cellulose pre- sents the characters of a dissolved compound of which the component groups are in the reactive state ; therefore that previous conversion into ' soluble ' products, as ordinarily understood, is not a necessary preliminary to destruction by the ferment organism. It is of interest in concluding this survey of research work, to reduce our conclusion to a short but comprehensive summary. Have we any alternative to offer to the views which will be General Summary 155 found formulated in the standard text-books ? It is, of course, generally true that research in any field impresses the worker with the inadequacy of antecedent accounts of any group of phenomena, such as are to be found in the text-books. The didactic method rather implies the legal formula of complete- ness — that is, it affects the statement of facts in the sense of 'the truth, the whole truth, and nothing but the truth.' And this appears to be justified in many sections of the science where the total synthesis of compounds is taken as the evidence of complete knowledge. In many provinces also, the develop- ment and refinement of experimental methods have enabled investigators so exhaustively to penetrate those relationships of form or structure, which we express in the term ' constitution,' as to beget the conviction that the formulae which express the relationships are a complete representation of the ultimate facts. We need only cite the evolution of ' stereology,' more especially that branch of it which has given us a complete account of the isomeric sugars — monoses — to show what we have to recognise in the general progress of exact knowledge. At the same time it is precisely in the same field that the limitations of this penetration into the ' actualities ' of matter are quickly apparent, for with the comprehensive knowledge of the ' con- stituents of cane sugar,' the constitution of cane sugar itself is still unsolved; and with the more complex members of the carbohydrate group, the problematical elements of ' constitu- tion ' multiply in proportion to complexity of proximate com- position. It is a fair inference that other constitutional factors are introduced in the equilibrium of these more complex members of the group, and have to be reckoned with in attacking these unsolved problems. We may presume that the first indica- tions will be furnished by the reactivities of the monoses in aqueous solution. We note the following points : 156 Researches on Cellulose, 11 1. The typical CO group is labile, and the presence of the alkaline hydrates — i.e. of OH ions — raises its ' mobility ' to the point that typical monoses are constitutionally interchange- able — e.g. dextrose, levulose, and mannose. 2. With increase of the OH ions and prolonged action there is a rupture of the Cg nucleus ; a complex of products result, the general feature of the resolution being an electrolytic migration. 3. With non-oxidising acids there is a series of decom- positions very complex in character, but with similar general characteristics. 4. Most characteristic, of course, are the decompositions determined by ferments, which are in an important sense a property of the carbohydrates in the sense of a function of their constitution. We must not overlook the fact that in these effects the influence of stereoisomerism or ' configura- tion ' are operative as modifying factors, but it is equally obvious that the general character of the decompositions is that of an interior electrolytic migration. 5. In various synthetical reactions with bases, and in crystallising with electrolytes, the monoses exhibit properties which are not revealed by their constitutional formulse ; though, of course, there is no inconsistency between such formulae and this group of reactions which they fail to indicate. If, then, the monoses in aqueous solution are electrolytes in the indicated sense, we have to inquire into the probable influence of this property in determining the constitution of the polyoses. It must be admitted that the present tendency is to assume for the entire carbohydrate group the comprehensive incidence of the molecular formulae — i.e. of the factors of constitution (and configuration) which they connote. In other words, that the 'polyoses,' including starch and cellulose, are in effect nothing more than multiples of the monose unit, these units General Summary 157 essentially persisting in the complex, which merely introduces the additional but subordinate factor of linking by oxygen bonds. This, however, is hypothesis current in default of any positive alternative ; hypothesis in the sense that it begs the entire question of the associated physical properties of these compounds. It assumes also that these ' organic bodies,' in the sense of vital products, retain no characteristic impress of the conditions under which they originated. In general opposition to this view, chemists will be quite prepared to take up at least the agnostic position, recognising the primary fact that no molecular formula is finally justified, except by such physical measurements as fix the dimensions of the unit. They will then admit that, to emerge from this position, a working hypothesis is necessary, and that that hypothesis is most acceptable which formulates the greatest number of experimental questions. This in our opinion is furnished by giving full value to the physical properties of starch and cellulose as typical colloids, and applying to their investigation the general considerations which apply to the entire group of colloids, inorganic as well as organic ; and chiefly that the colloidal state is not merely the accidental or incidental attribute of 'large molecules,' but a condition of matter determined by chemical factors — that is, by chemical functions, by valency and intrinsic energy or potential. Thus we have found in the course of the present inquiry : 1. That the solid state of cellulose is continuous with that of its solution ; there is no break in passing from one to the other, as in the case of compounds which crystallise. 2. Cellulose, in presence of water, is an electrolyte ; it passes the current without decomposition, undergoes interior changes, and manifests a ' polar ' activity. It does not revert to the original condition when the current ceases to pass. We have hence defined its electrolytic characteristics as 'intrinsic' 158 Researches on Cellulose, II 3. The presence of active ions in the fibre colloids, of which cellulose is a prototype, affords a consistent explanation of the phenomena of dyeing, involving a ' coagulation ' stage, by interaction of ions of opposite electro-chemical function ; and a stage of osmotic penetration of the fibre colloid, also facilitated by differences of .potential, and in many cases attended by demonstrable ionisation of the colouring matter taken up. 4. The solutions of cellulose in aqueous solvents are the result of its amphoteric constitution ; the solutions result from the reciprocal interaction of groups of opposite electro- chemical function. 5. Cellulose in contact with H and OH ions is continuously modified, both in chemical function and in structure ; the modification is a function of the intensity of the action and its duration. Even when insoluble in the solution of the electrolyte, the cellulose is, nevertheless, in ' solution contact ' with the active groups. The interpenetration is of the nature of that of two solutions perfectly miscible. 6. Cellulose is resolved by specific ferment actions. 7. Cellulose reacts synthetically to form esters and the reactions have continuous quantitative characteristics. There are no breaks marking ' molecular ' stages, and the products which may be and doubtless always are mixtures, are not re- solved by solvent fractionations as in the case of crystalline derivatives. The limiting stages correspond with the stoichio- metrical ratios of the interacting unit groups, and vary with the chemical functions of the combining negative groups. Cellulose in these relationships is not less, and in some cases is demonstrably more, reactive than the simpler carbohydrates. In these reactions it maintains its structural characteristics unaltered, or, if previously modified, as by conversion into hydrocellulose, the modification persists in the ester stage. General Summary 159 By saponification or spontaneous decomposition there is reversion to ' cellulose,' and in many cases the cycle, though marked by reactions of considerable intensity, is completed with only fractional changes of weight. These are the more important facts to be summarised in a , constitutional formula, if such were possible. If we revert for a moment to the consideration of the con- ditions of origin of cellulose, we must conclude that the plant- cell has not synthetised monose groups as such into a complex multiple or polymeride. Having regard to the extreme sensi- tiveness of the monoses in aqueous solution, it is not to be supposed that they remain unresolved within the sphere of activity of the cell. On the other hand, the fact that the cellu- loses are resolved by specific acid treatments to monoses would imply that the actual component groups are ionised residues of the monoses. This view is not without its a posteriori evi- dences ; these have indeed been categorically stated in the preceding pages. It is considerably strengthened by the in- clusion of cellulose in the arena of investigation of the colloids as a class, by regarding cellulose as a typical representative of the colloidal state of matter and its properties in reaction as depending upon physical conditions specifically different, though not distinct, from those which determine crystallisation. If this is conceded, two consequences result. First, that molecular constitutional formulae are inapplicable to the celluloses ; secondly, that progressive research must be based upon dynamic methods of the general type applicable to solutions. As workers in this field we may say that these alternative views continually provide consistent explanations of pheno- mena which otherwise would be referred to the unrelated idiosyncrasy of a chemical individual; but, more important still, they have suggested numerous directions of fruitful i6o Researches on Cellulose, II investigation. For the present we must be content to develop the experimental subject-matter on the lines of these later re- searches, in close association with the most advanced move- ments of molecular science. The purpose of this essay is to aid in putting the subjects in this, which we think will be recog- nised as a more comprehensive perspective. SECTION III TECHNICAL PROGRESS The period 1900-1905 has been marked by considerable activity in the development of the cellulose industries ; but we are not aware of any movement which can rank as a ' de- parture ' in this branch of technology. In the long-established staple industries involving chemical processes of preparation or treatment, such as textile bleach- ing, dyeing and printing, paper-making and the preparation of paper-makers' raw materials (celluloses), the only noteworthy feature is the general expansion and the increased perfection of method. In the wet spinning of short cellulose fibres by the processes known as the Kellner-Turk and Kron systems respectively, there has been a notable industrial development, and the present output of these cellulose yarns is estimated to be from 5 to 10 tons per day. In the 'artificial silk' or ' lustracellulose ' industry, the developments have been phenomenal. In the year 1899 the product was still regarded as an industrial curiosity, with a probable if problematical future. In 1900 it had vindicated for itself an unassailable position as a textile, and in the pre- sent year (1906) it has reached such a position that the esti- mated daily output in Europe is 7,000 kilos., increasing to 9,000 by the end of the year. A noteworthy feature of this advance in production is the correlative fall in selling-price, which may be estimated, on mean figures, at from 35 to 20 francs per kilo., say from 13J. to 7 J. per lb. In the present deve- 1 62 Researches on Cellulose, II lopments three systems are competing, the collodion process (Chardonnet-Lehner), the cuprammonium process (Pauly), and the viscose process (Cross and Bevan-Stearn-Topham). The final industrial equilibrium as between these systems on the one side, and the adjustment as between these artificial silks and the natural silk on the other, is a subject for speculative opinion, as well as for a judgment based upon the ultimate factors which are technical — that is, scientific. It is to be noted that the demands made upon the artificial silks in point of fineness and uniformity are increasing in severity, and a large percentage of the present production is within the limits of fineness represented by a thread of i lo to 130 deniers. This is already a large advance upon the range of 160 to 220 deniers, which was current a year or two ago. It should be borne in mind, however, that these numbers represent aggre- gates of filaments. If we consider the actual unit filament, the artificial silks are produced at an average range of 5 to 8 deniers, which is about twice the weight of the true silk as spun by the silkworm. In the production and applications of cellulose acetates there has been a noteworthy advance. There are now three centres of this manufacture, in the hands respectively of ' Fiirst V. Donnersmarck Kunstseide und Acetatwerke,' the well-known manufacturing corporation ' Farbwerke vormals F. Bayer & Co.,' and A. D. Little, of Boston, Mass., with his associates Walker and Mork. One of the most important uses of these products is in insulating fine copper wire of dimensions 0*07 mm. to o-i7 mm., by applying a coating of the acetate of about 0*02 mm. thickness. This wire has not only high insulating capacity, but an extremely low specific induc- tion, and is being extensively used in the electrical industries. Other important applications of this group of esters are awaiting the necessary reduction in cost of production. Technical Progress 163 Applications of Yiscose. We have already spoken of the conversion of ' viscose ' into a lustracellulose or ' artificial silk,' but this is only a particular effect. The applications of the cellulose sulphocarbonate, in fact, are many sided. They are necessarily fraught with the obvious difficulties determined by the characteristics of these compounds. lis ont les defauts de leurs gualites. To this cause must be ascribed the tardy development of the special industrial effects and products which are based upon the applications of ' viscose.' It is important to take these in brief review, inot only on account of the special technical effects, but as a document illustrating some conditions of technical progress which are not generally taken into account. The following products are now manufactured in increasing quantity : Paper and Boards from pulps sized with viscose in the engine. Various particular effects are compassed of which the more important are : high strength and stretch, resistance to water, surface finish, generally the appreciation of raw materials with less preparation (beating). Incidentally should be noted the colour-fixing function of the regenerated cellulose ; a large number of ' engine ' colourings are very much faster to light in the case of viscose-sized papers. The surface and ' body ' sizing of papers by means of viscose produces a number of effects which may be deduced frdm the quite exceptional properties of cellulose as a plastic colloid, the structural properties of which are uniformly continuous. But the special results obtainable have only been industrially developed in one direction. This applica- tion, however, involves the general technical principle of ' appreciation.' It is evident that a paper sized with ' viscose ' — M 2 164 Researches on Cellulose, 11 i.e. indurated with the regenerated cellulose — is able to resist the action of water and boiling solutions, and it can therefore be dyed, printed, and otherwise handled as a textile fabric. This is at once a considerable extension of technical practice and method, and a widening of the province of application of papers. Papers treated in this way, dyed in the ordi- nary 'jigger' machine, printed by calico-printing methods, embossed and otherwise ' finished ' by the method applied to textile cloths, are now produced on the industrial scale and will in due time take their place as articles of general usefulness. Applications to Textiles. — The uses of viscose in the dyeing, printing, and finishing of cellulose textiles have been very slowly developed. Certain special effects were early recognised as producible by means of this product — e.g. fast pigment prints upon calico, lustre finishes on cotton goods, and ' shrunk ' finishes on unbleached flax and hemp goods — and these have been steadily developed. More recently special attention has been given to black lustre finishes which are now industrially worked. But the most interesting and important applications are those which are more obviously conditioned by the special properties of cellulose, in which the proportions used — i.e. added — are sufficiently large to give particular effect to them, of which we may mention : 1. Finishes upon cloth in which the cellulose takes the form of a continuous film. Such ' finishes ' are applicable as book cloths, window-blind cloths. They are characterised necessarily by a considerable addition of substance and weight, — e.g. 20 to 50 p.ct. — and the added substance being a true 'cellulose' constitutes an obvious superiority to starch, gelatin, and other filling agents previously employed. 2. Textile yarns, chiefly cotton and flax, are coated with viscose, which is applied and coagulated in sitH by a special mechanical process, afterwards fixed, and the yarn finished and Technical Progress 165 dried, so as to conserve the lustre tenacity and elasticity of the structureless cellulose. The addition of cellulose is carried to three times the weight of the original textile, and with these large proportions the original yarn is so entirely enveloped, that its presence can only be recognised on close examination. The products have somewhat the appearance of a horsehair, but of regular dimensions and of continuous length ; they are, however, more lustrous, and the effects obtained in the goods woven with them may be described as an unusual combination of firmness to rigidity with a certain softness of touch and a silky lustre. These products are now making rapid headway on the Continent,, in the hands of the exclusive licencees under the respective patents — in France, the Societe Frangaise de la Viscose, who were the originators of these developments ; in Germany, the Fiirst v. Donnersmarck Kunstseide und Acetat- werke. Cellulose Films. — In the effects above described, the added cellulose constitutes a film or visibly continuous mass. Many and exceptional difficulties have stood in the way of pre- paring the cellulose — regenerated from viscose — as a film, in continuous length and large dimensions, without the support of a textile fabric. But the pure cellulose, in the form of a transparent film, is now industrially produced (a) as a tube for use as a dialysing septum, and (V) a capsule for sealing bottles containing 'chemical,' pharmaceutical, and dietetic prepara- tions. The former have been adopted after lengthened experi- mental trials at such important centres of work and investiga- tion as the Pasteur Institute. No publication has been made of the results obtained with these tubes as a contribution to the theory and practice of osmotic separations. We can therefore only reproduce the statement in general terms of the particular efficiency of cellulose (hydrate) as a homogeneous 1 66 Researches on Cellulose, 11 membrane. It has obvious specific advantages in many cases as a non-nitrogenous medium, and through being particularly resistant to the actions of alkalies. It has advantages over parchmentised cellulose, not only from the fact of its unifor- mity of composition, but the certain absence of 'pin-holes' — in short, by reason of its structural perfection. The use of viscose-cellulose capsules is an interesting application of properties which in other directions constitute the particular impediment barring its employment. The viscose being applied to the surface of a cylinder with rounded end, or tube, is ' precipitated ' in sM and ' fixed,' and is then a tube of hydrated cellulose which is detached and washed. On exposure to the air the hydrate parts with its water, and as water of hydration constitutes 85 p.ct. of the original mass, the resultant shrinkage is correspondingly large. In shrinking upon any surface such as the neck of a bottle, the cellulose attaches itself automatically at all possible points of contact and constitutes a complete seal. Cellulose in Massive Forms ; Viscoid and Viscoid Agglomerates. — Viscose by its spontaneous reversion to ' cellulose ' forms masses of hydrated cellulose of the exact dimensions of any containing vessel ; the process of shrinkage then continues, and under conditions which maintain a uniform process of dehydration the cellulose becomes a uniformly com- pacted mass of extraordinary hardness. In the preparation of industrial products the dehydration is necessarily preceded by an exhaustive process of washing to remove the by-producta of the interaction of the alkali and sulphur residues. The cellulose may also be dyed or coloured, and an indefinitely varied range of effects are then produced. These have been specially developed in the hands of the Societe Frangaise, and the Viscolith which they produce is an ornamental structural product having a number of special applications. Technical Progress 167 In admixture with various waste products, organic and inorganic, viscose constitutes an agglomerating or compacting medium, and a number of artificial structural solids may be produced. A black agglomerate composed of cellulose, china clay, or other moist filling material, with pitch and similar dense hydrocarbons, is prepared by milling the viscose with these ingredients, forming the plastic mass into cylinders or cubes, 'setting,' washing, and dehydrating the washed solid. The product is used in the engineering and electrical indus- tries, and is known as ' viscoid.' Viscose and Indiarubber. — Viscose, as indicated above, can be milled with hydrocarbons to a species of permanent emulsion ; and this property is taken advantage of in the pro- duction of a number of new technical effects in the indiarubber industry. The admixture of the cellulose with this group of hydrocarbons is effected in various ways, and the products differ from the ' viscoid ' described in the preceding section, chiefly in the added process and effect of vulcanisation. The cellulose is similarly used in producing film effects (water- proofing) and in the solid agglomerates known as ' mechanicals. This outline of the many applications to which these deriva- tives of cellulose are adapted is to be taken as an illustration of practical results deducible from the properties of cellulose. In these applications only those effects are evident which belong to cellulose as a typical colloid. But in the inter- mediate, soluble, and plastic stage, the special chemistry of cellulose is brought into evidence. The complications attending the introduction of alkali and sulphur residues are very serious ; and further difficulties result from the inherent tendencies of the products to spontaneous decomposition. The control of a complex equilibrium to which so many 1 68 Researches on Cellulose, II factors contribute is a task which imposes very methodical and exact working. It is not to be expected, therefore, that manufacturers will devote their resources to the solution of the problems of practical working unless impressed with the conviction of certain success, and the value of the results when success is attained. We are of opinion that the applica- tions of viscose are only a special case of the general difficulty of manipulating colloids in the absence of a general theory of the colloidal state, and the advantage of a theory or a working hypothesis is that it aids in fixing those convictions which precede successful enterprises. The industrial applica- tions of the colloids have, it is true, been largely developed on a purely empirical basis. We look forward to considerable expansion of these industries with the adoption of a general theory.^ A general survey of cellulose technology produces the impression, as we have often had occasion to remark, of a high state of development of these arts, marked by extraordinary refinement of method and an accumulation of exact know- ledge which, however, would scarcely rank as ' exact ' by comparison with other branches of the science, or in the usual sense of the term as applied to scientific knowledge. We think this arises from the want of a theory, or generalisation ap- plicable to the cellulose groups, as to compounds in being — i.e. in action and reaction. It is not necessary in dealing with the carbon compounds generally to separate this, the phase of actuality, from the questions belonging to their synthesis and decompositions. To put this point shortly, a vast majority of the carbon compounds may be accounted for in terms of their '■ An important publication giving an exhaustive review of progress in this field is Die Viskose, B. M. Margosches, Leipzig, 1906 (Klepzig), a brochure of 130 pages and supplement dealing with bibliography and patent literature. Technical Progress 169 past and future — i.e. their origin and decomposition — without reference to their present condition. So we may endeavour to form a mental picture of cellulose by studying its products of final proxirnate resolution, and as we cannot study the synthetic process under which the constituent groups are put together, we draw on our imagination, with the assistance of numerous analogies, and, assuming that they are linked by the simplest constitutional bonds, we postulate the molecular aggregate, and admit a doubt only as to the exact form of this linking, and the dimensions of the molecular unit which results. Now we have endeavoured to show in our exposi of recent researches and contributions to the science of the subject that this method of dealing with ' cellulose ' as a problem in consti- tution, when the colloidal state is introduced as a factor, necessarily breaks down. All we know of the products of resolution, and of the ultimate reacting groups of cellulose, tells us nothing of its origin and of the conditions under which it is synthesised, and throws only a piartial light on its actual internal conditions as a body in being. We have endeavoured to indicate the directions of evolution of more and more com- prehensive formulae ; and have also suggested that the value of a working hypothesis will be measured by its providing new directions of attack. It will aid to establish this new position of critical study if we indicate some of the avenues of investi- gation through which it may be applied to technological problems. We select a few typical cases : PapeF-making. — The process of beating or wet-milling of paper pulps is usually regarded as a ' mechanical ' treat- ment : a comminution by reduction of the fibres to the neces- sary condition of division. But on our present view it is the milling together of two solutions. These do not develop opposing surface tensions, though the cellulose exhibits a 170 Researches on Cellulose, II certain resistance to penetration by water. In recent years attention has been devoted to the effects of prolonged beating, the striking result of which treatment is to bring the cellulose to the condition of a colloidal solution ; so much so that the cellulose possesses the essential physical characteristics which it manifests in solutions obtained by chemical means. The differences are those of degree. The investigation of the process of hydration becomes a problem in 'surface energy.' It may involve the question of a difference of potential between the cellulose and the water. It will certainly be effected by the presence of electrolytes, and probably favour- ably influenced by amphoteric electrolytes. The ordinary processes of beating for the preparation of pulp for the paper machine range in duration from 20 minutes to 14 hours, and although not directly aimed to compass these effects of colloidal hydration, they and their attendant phenomena are of course involved and should be specifically studied. The complex effects produced in the process of 'engine sizing,' which are usually referred to the intrinsic effects of the added sizing agents, rosin (soaps), starch, sulphate of alumina, obviously involve a web of colloidal interactions, and it cannot be a matter of indifference to the technologist as to how these interactions succeed one another, and what is the ultimate' balance of effect or equilibrium. To investigate and control these interactions we need to apply the methods and theories which are contributing to the rapid advance of our knowledge of the colloidal state. Textile Bleaching and Finishing. — The chemical pro- cesses employed in preparing the cellulose textiles for their ultimate marketable condition are usually regarded merely as methods of eliminating ' impurities,' coloured and otherwise. But it is obvious that the cellulose itself is modified in interior Technical Progress 171 constitution by every treatment to which it is subjected. These modifications are, for the most part, entirely disre- garded ; in fact, for purely textile uses the tests applied are entirely external in significance. It is true, however, that certain directions of employment impose conditions which involve altogether different factors, and these have made technologists aware of a wide range of variations in celluloses, which would rank as of an invariable standard if judged from the standpoint of structure and external appearance. We may instance more especially the celluloses required for the preparation of nitrates, acetates, sulphocarbonates, and the aqueous solutions of cellulose. In these directions of employ- ment the interior constitutional factors manifest their influence. Not only is the course of the reactions of synthesis effected, but the physical properties of the final product. As a single illustration we may take two samples of cotton, one of which has been ' modified ' by the process of mercerisation, and expose them to esterifying conditions, for the production of nitrate or acetate. In the one case we obtain no reaction, or only such degree of reaction as indicated by a gain of weight of 6 to 10 p.ct., whereas the original cotton under the same con- ditions gives a maximum reaction with a gain of weight of 70 to 80 p.ct. The technology of the nitrates of cellulose, while it im- poses a very exact study of the raw materials employed, has not contributed to the scientific co-ordination of these facts, as might have been expected, probably from the reason that the knowledge is withheld from publication. In any case, however, the basis of this knowledge is empirical, so far as it is accumulated without the direction of a working hypothesis as to the constitution of cellulose. This hypothesis we have now suggested ; it will be found to open up a number of lines of investigation which, as they have already been found to be 172 Researches on Cellulose, 11 fruitful, will no doubt contribute to more general progress in this branch of technology. But if these synthetical reactions of cellulose are affected by the antecedent conditions of its preparation, reciprocally, these reactions can be employed as a means, certainly the most important means, of diagnosing the changes determined by the processes employed by the bleacher. This opens a wide field for investigation, together with the prospect of co- ordinating the unclassified experiences of technologists in this branch with a general theory of cellulose constitution. The operations of ' finishing,' which also appear to be hedged round with a great deal of the ' secrecy ' which appertains to the ' fine art ' sections of the cellulose industries, are capable of many explanations deducible from the general views herein expounded. Every operation through which a cloth passes leaves its impress on the constitution of the cellulose, and therefore on the factors which determine structural qualities, and even such external features as ' surface ' and ' lustre ' and feel. The essential factors of finishes are, therefore, the actual constitution of the cellulose unit ; and the modifications de- termined by added ' conditioning ' agents are of secondary importance. The relation of ' finishing ' quality to the bleaching pro- cesses employed becomes, on this view, a definite object of investigation. Probably it will be found that the most im- mediate criteria of the condition of a bleached textile is its relation to atmospheric moisture and its specific afiSnities to typical dyestuffs. Underlying these relationships are the more fundamental conditions of constitution, which can only be brought to light by applying the test of a synthetical reaction. In any case, from our present point of view it is not satisfactory to judge the quality of a ' bleach ' merely on the external evidence of colour and textile quality. Progress Technical Progress 173 must result from an exercise of our grasp of the unseen and underlying causes of quality. Dyeing and Printing. — ^What we have indicated above is already in part recognised in relation to the routine practice of dyeing and printing. As an instance, a ' printer's bleach' of cotton, cloth, or calico is a different effect from that which fulfils the standard of a ' market bleach,' and is tested by a special dyeing test in an alizarin bath, followed by soaping. The ' purity ' of the ' bleach ' is inversely as the depth of tint communicated to the cellulose. This is an empirical test, but one which, in correlation with others such as we have indicated, may be made to furnish much fuller information as to the actual condition of the cellulose. But it is in relation to dyeing phenomena that we have been able to show the gradual recognition of the amphoteric activity of cellulose ; more than that, that as a typical colloid its be- haviour with colouring matters depends in the first instance upon coagulation by interactions of ions of electrically opposite function, ions of the colouring matter and ions of the fibre colloid. This proceeds from a positive consistent theory, or at least a working hypothesis, in which are merged and generalised the older partial theories represented by such terms as ' mechanical ' and ' chemical ' applied to particular modes of explaining these phenomena. We are still of opinion that there has been a disparity of the attention of investigators in regard to the two sides of dyeing phenomena; the fibre colloids have been relatively little studied. With a general hypothesis as to their constitu- tion in the form of a positive connecting-link with the modern ■developments, of the ' theory of solution,' the way is opened for definite experimental issues to be put and solved. Artificial forms of Cellulose and Cellulose Deriva- tiYes. — It is necessarily to this branch of technology, which 174 Researches on Cellulose, II includes the industries in artificial silk (lustracellulose), monofil- fibres and solid agglomerates, that we look for the most impor- tant contributions to the general theory of the soUd state, and to the particular problems of cellulose constitution. It will be appreciated that the industrial process of con- verting the fibrous celluloses into 'artificial silk' involves an extraordinary refinement in the adjustment of factors of essentially variable character, and many of these variables are intimately connected with the constitution of the cellulose employed. It is equally evident that these industries have been perfected upon a vast accumulation of the results of exact experiments, which are scientific documents of consider- able prospective value when applied in connection with general theories ; both in verification of these theories and by re- ciprocal contribution of new matter for the extension of these theories. Thus these processes give us control of the plastic forms of the celluloses and derivatives, and we can produce them in regular continuous form, or solid aggregates of any desired dimensions. This brings a number of the physical and mechanical properties of the celluloses within the scope of exact measurement, seeing that the factors of surface and volume can be determined within the limits of ascertainable error. As we have frequently pointed out, these physical pro- perties are specifically connected with chemical constitution, and the study of these relations is much facilitated by the con- trol of factors which vary so irregularly in the ' natural ' forms^ of cellulose as to make exact determinations impossible. It is not the purpose of this essay to formulate the many objectives of research in terms, nor to anticipate the results of such research. Our aim is to provoke discussion and generally to indicate directions of investigation, which we hope will be undertaken without regard to ' specialist's ' qualifications. ADDITIONAL BIBLIOGRAPHICAL NOTES We briefly notice certain publications which have important bearings, direct or indirect, upon the subject-matter of this volume, without however contributing to the immediate eluci- dation of the special points discussed. OBSERVATIONS ON COTTON AND NITRATED COTTON. H. DE MosENTHAL (J. Soc. Chem. Ind. 1904, 23). Contains the results of laborious observations of the minute structure of the cotton fibre and of the derivative esters, with an examination of the characteristics of their solutions, including observations on the dialysis of solutions of the nitrates. The author's methods and results, when extended and developed, will be a valuable contribution to the solution of ■ the problem of the relation of visible structure to the constitu- tion of the aggregate or reacting unit. THE SURFACE STRUCTURE OF SOLIDS. G. T. Beilby, Third Hurter Memorial Lecture (J. Soc. Chem. Ind. 1903). This is a contribution to the general problem of solid structures, but specially directed to the condition of matter in 1/6 Researches on Cellulose, II metallic masses, the surfeces of which are composed of films, with a proportion of spicular masses of minute dimensions which have characteristics sufficiently invariable to be regarded as structural components. The film-matrix is the evidence of the tendency to surface-flow under the action of pressure, •which the author has observed to occur not only in metals but in the widest range of solid bodies examined, 'glass, agate . . . salts of all kinds . . . gelatin and cellulose.' C.B.S. UNITS AND STANDARD PAPER-TESTS: AN ESSAY TOWARDS ESTABLISHING A NORMAL SYSTEM OF PAPER-TESTING. C. F. Cross, E. J. Bevan, Clayton Beadle, and R. W. SiNDALL (London, 1903, Wood Pulp, Ltd.). This is a symposium, the purpose of which is sufficiently •explained by the title. It is pointed out that the standards adopted in Germany and on the Continent, with their obvious advantages, are defective in important respects. The con- ventional standard of ' breaking length ' introduces certain errors which in the case of loaded papers may be very large, and which are to the advantage of the buyer as against the • paper-maker. The authors point out that the numerical records of use to the paper-maker in controlling his manu- facture are those which set out the physical characteristics of papers in the simplest terms, and they give particular pro- minence to measurement of air-space (as a measure of ' bulk '), breaking-strain (a) per unit of total sectional area; (b) per unit of area of fibrous components (' specific tenacity '). Additional Bibliographical Notes 177 BIBLIOGRAPHIE DER KOLLOIDE. Arthur Muller (Leipzig, 1904, L. Voss). This bibliographic resume of investigations of. colloids and the colloidal state is of particular value at the present juncture, when the subject is attracting attention in so many directions. The author notices more than 350 original publications, and covers a much more extensive range than that comprised in the ' Index to the Literature of Colloids ' of Whitney and Ober (J. Amer. Chem. Soc. 1901, 23, 856-63). It is perhaps of interest to note that ' cellulose ' is con- spicuous by its absence from the extensive range of subjects and objects investigated. PAPIERSTOFFGARNE : IHRE HERSTELLUNG, EIGEN- SCHAFTEN UND VERWENDBARKEIT. By E. Pfuhl (Riga, 1904, G. LofHer). This is an important monograph of the new industrial developments based on the wet-spinning of short fibres. Paper-makers' ' pulps ' are reduced in the beating engine, run on a paper machine into strips in continuous lengths, which are afterwards subjected to a twisting operation and converted into yarns available for weaving. Two or three systems are now commercially developed and the industry is of growing importance. The author treats fully of this aspect of the matter, while devoting considerable attention to the technical characteristics of the products. It is an interesting contribution to the general question of the structure of fibrous aggregates and their derived physical or mechanical properties. The work comprises one hundred and fifty pages, with numerous tables and illustrations of the machinery employed. 11. N 178 Researches on Cellulose, II ON THE ACTION OF WOOD ON A PHOTOGRAPHIC PLATE IN THE DARK. W. J. Russell (Phil. Trans. B. 1904, 197, 281). This is a further contribution of experimental observations of the curious chemical activity of wood surfaces, brought to demonstration by actions on sensitised plates in the dark. These actions are influenced by previous insolation of the objects. Exposure to light also renders active various resins, turpentine, and varnishes. These observations are of con- siderable prospective interest, but in the absence of any positive conclusions of the author as to the actual cause of these effects, we can only note them as an index of an active condition of ligriocellulose surfaces. cellulose:, CELLULOSEPRODUKTE und KAUTSCHUKSURROGATE. Joseph Bersch (Leipsig, 1904, A. Hartleben). Authorised English translation of the above, by W. T. Braunt, published under the title, ' Cellulose, Cellulose Pro- ducts, and Artificial Rubber,' London, 1904, Kegan Paul, Trench, Triibner & Co. The scope of this work is in the main ' technical ' ; it aims to give a full account of the more recently developed cellulose industries. The following heads of chapters will give a general idea of the work : (i) ' Cellulose,' a general sketch of the chemistry of cellulose ; (2) Wood pulp (' mechanical ') ; (3) Wood pulp ('chemical'); (4) Vegetable parchment; (5) Manufacture of sugar and alcohol from wood and wood cellulose ; (6) Manufacture of oxalic acid from wood ; (7) Viscose and viscoid; (8) Nitrocellulose (gun-cotton, pyroxylin); (9) Artificial silk; (10) Artificial cellulose fibres ; Additional Bibliographical Notes 179 (11) Viscose-spinning; (12) Celluloid; (13-14) Indiarubber substitutes. The matter is comprised in four hundred pages, with a full index. VISCOSE : PREPARATION, PROPERTIES, AND APPLI- CATIONS, WITH SPECIAL REFERENCE TO ITS USES IN THE TEXTILE INDUSTRIES. B. M. Margosches (Leipzig, 1906, A, Klepzig). Die Viskose — ihre Herstellung, Eigenschaften und Anwendung. This is an exhaustive account of the technology of the cellulose xanthogenic acids, the descriptive portion occupying a hundred and thirty pages of matter, and a bibliography of the subject forming an appendix of eighty-seven additional pages. The substance of the work appeared in a series of articles in the Zeitschrift f. d. Gesamte Textil- Industrie, during the period 1904-06. The information contained in this work is not only extensive, but — what is by no means always to be said of works of this character — correct. The encyclopaedic character of the book conforms with its obvious purpose and scope, and precludes a more extensive notice. It will be recognised as a valuable addition to the literature of ' cellulose,' and indispensable to those who are interested in technological developments. CHAPTERS ON PAPER-MAKING. Clayton Beadle (Vol. L 1904; Vol. II. 1906; H. H. Grattan, London). The first of these volumes reproduces a series of ten lectures delivered by the author on behalf of the Battersea Polytechnic Institute, and traverses a number of subjects of l8o Researches on Cellulose, II special interest to paper-makers desirous of following up the scientific technology of the manufacture, including ' Raw Materials,' 'Bleaching,' 'Physical Properties of Papers and Paper-testing,' ' Chemical Residues in Papers.' The treatment of these subjects is original, and the author embodies the results of several of his own experimental investigations. The second volume is more ' educational ' in its purpose and scope, and the author has devoted himself to a full consideration of the examination papers of the City and Guilds Institute, on the subject of paper-making, during the period 1901-05. INDEX OF AUTHORS Beadle, C, 179 Beadle, C, and Dahl, O. W., 70, 74 Bebie, 30 Beilby, G. T., 175 Beinadou, J. B., 66 Bersch, J., 178 Bevan, E. J., see Cross Biltz, W., 120 Bumcke and Wolfen- stein, 144 Campbell, A., 130 Cross, C. F., and Bevan, E. J., 93, "7. 133. i39> 176 Cross, Bevan, Beadle, and Sindall, 176 Cross, Bevan, and Briggs, J. F., 83 Cross, Bevan, and Jenks, R. L., 51 Darling, C. R., 128 Dreaper, W. P., 123 Geinsbergen, E., 144 Girard, A., 138 Green, A. G., 131, 136 Hake, ■ C. W., and Lewis, R., 53 Hiibner, J., and Pope, W. J., 106, 113 JoRDis, E., 6, 126 Kellner, O., 145 Knaffe, E. von, 144 Knecht, E., 48 Lunge, G., 30 Macfadven, a., and Blaxall, F. R., 153 Masson, O., 25, 70 Margosches,' B. M., 168, 179 Meldola, R., 3 Menter, F., 143 Mosenthal, H. de, 175 Miiller, A., 177 Nastukofp, a., 144 Omelianski, W., 151, Pfuhl, E., 23, 177 Pope, W. J., see Hiib- ner Russell, W. J., 178 SiRK, H., 144 Skraup, Z. H., 143 Stern, A. L., 138 Suida, W., 121, 122 Weintraub, 30 Will, W., 56, 73 INDEX OF SUBJECTS Acetates of cellulose, ii, 132, 162 Acetic acid from cellulose, 153 Acetochloiides of starch, glycogen, and cellulose (derivatives), 143 Acetosulphates of cellulose, 12, 83, 134 Acidification of cellulose, 104 Acids and alkalis, action on cellu- lose, 15, 132, 141 Alkali-cellulose, 13, 104 Amphoteric theory of cellulose con- stitution, 66, 102, 158 Artificial horsehair (viscose), 165 — silk, see Lustracellulose Assimilation of straw cellulose by herbivorous animals, 145 Benzoates of cellulose, 13, 94 Bleaching of textiles, 172 Bottle-capsules (viscose), 166 Brom-methyl furfural, 15, 132 Butyric acid from cellulose, 153 Cathode image, latent, 128 Cellulose, acetates of, 11, 132, 162 — acetochloride, 144 — acetosulphates, 12, 83, 134 — action of acids and alkalis on, 15, 104, 132, 141 — action of ferments on, 16 -^ action of oxidising agents on, 16 — as an electrolyte, 7, 103, 128, 130. 157 — as a solution-aggregate, 7, 159, 169 Cellulose, assimilation of straw, 145 — benzoates, 13, 94 — constitution, amphoteric theory of, 66, 102, 158 — constitution and physical struc- ture, 4, 28, 69, 135 — constitutional formula (Green's), 133 — decompositions, theory of, 17 — degrees of nitration of, 32 — destructive distillation of, 17 — determination of unchanged, in nitrocellulose, 31 — esterification, theory of 'con- tinuous,' 47, 100, 158 — esters, dyeing of, 122 mixed nitric and sulphuric, 51,52 — fermentation of, 151 — hygroscopicity of, 73 — labile nitrate of, 48 — molecular weight of, 144 — nitrates, 11; see aXso Nitrocellu- lose — products, 178 — solvents of, 11 — specific heat of, 71 — specific inductive capacity of, 130 — sulphates, 12 — systematic description of, 10 — xanthates, 14, 95, 163 — xanthogenic acid, 93 Celluloses, groups of, 2, 18 Colloids, bibliography of, 177 — coloured inorganic, 120 — theory of, 6, 126 Cotton, absorption of moisture by, 27,70 Index of Subjects 183 Cotton, absorption of moisture by, theimal effects of, 71 — different grades of, for collodion, 44 Cotton yarns, influence of reagents on, 106 relations of strength and twist of, 108 mercerised ; lustre, tinctorial, and structural properties of, 113 Decomposition reactions of cellu- lose, theory of, 17 Destructive distillation, 17 Dialysing tubes (viscose), 165 Dyeing, colloidal factor in, 126, 158 — of textiles, 173 — theories of (Cross), 117, (Biltz) 120, (Dreaper) 123 DyestufTs, affinity of mercerised cotton for direct, 114 — in relation to chemical functions, 122 — on starch, silica, and silicates, 121 Electrograph, 128 Electrolytic properties of cellulose, 7, 103, 128, 131, 157 Fatty acids from cellulose, 153 Fermentation of cellulose, 151, 153 Ferments, action of, on cellulose, 16 Ferricferricyanide reaction, 125 Ferric hydroxide hydrosol, 6 Fibres, dimensions of, 20 — spinning of textile, 21 Fodder experiments, 146 Glycogen acetochloride, 144 — molecular weight of, 144 Hemicelluloses, 19, 142 ' Horsehair,' artificial (viscose), 165 Hydrobromic acid, action of, on cellulose, 15 Hydrocellulose, 15, 13^, 139 Hydrogen fermentation of cellulose, 153 Hydrosols and hydrogels, 6 Hygroscopic moisture, effect of, on mechanical properties, 25 Hygroscopicity of cellulose and nitrocellulose, 73, 80 LiQNoCELLULOSES, resolution of, 149 Lustracellulose, 21, 161, 174 — dimensions, 22, 162 — mechanical properties, 23 effects of moisture on, 25 Lustre of mercerised cotton, 113 Mercerisation, 104, no — effect of, on lustre, tinctorial, and structural properties, 113 — shrinkage due to, 115 Methane fermentation of cellulose, 153 Moisture, absorption of, by cotton, 26, 70 — effect of, on mechanical pro- perties, 25 Monose sugars, electrolytic charac- ter of, 156 Nitrate of cellulose, a labile, 48 Nitrated cotton, 175 Nitrating acids, influence of com- position of, 39 temperature of, 37 water in, 36 Nitration, degrees of, 32 — influence of time on, 38 Nitric acid, absorption of, by cellu- lose, 49 Nitrocellulose, 11 — action of solvents on, 41 — causes of instability of, 52, 53 — hygroscopicity (Will's .re- searches), 73, 77, 80 — Lunge's researches on, 30, 45 — oxycellulose in, 43 — stability of (Will's researches), 56,61 1 84 Researches on Cellulose, II Nitro-sulphuric esters of cellulose, 51. 53. 90 Oxidising agents, action of, on cellulose, 16 Oxycellulose, 16, 43, 132 Paper, mechanical properties of, -^ structure of, 23 — testing, ' C.B.S. units," 176 — viscose sizing of, 163 Papermaking (beating), 169, 179 Paper-pulp yarns, 24, 161, 177 Photographic action of wood, 178 Physical structure and chemical constitution, 4, 28, 69, 135 Primuline, 124 Printing by electricity, 128 ' PyroceUuIose,' 66 Silica and silicates, dyeing of, 121 Silicate hydrosols, 6 Silk, artificial, see Lustracellulose Smokeless powders, 66 Solids, surface structure of, 175 Stability test for nitrocellulose (Will's), 59 Starch acetochloride, 143 — dyeing of, 121 — molecular weight of, 144 Straw-cellulose as fodder, 145 Sulphuric-acetic esters of cellulose, 83 Sulphuric acid, action of, on cellu- lose, 12, 15, 132 Sulphuric-nitric esters of cellulose, 51, 53. 90 Textiles, application of viscose to, 164 — bleaching and finishing of, 171 — dyeing and printing of, 173 Textile yarns, mechanical properties of, 23 paper-pulp for, 24 spinning of, 21 strength and dyeing pro- perties of, 106, 113 Thermal effect of wetting cotton, 71 Thermophilic bacilli, 153 ' Vital ' products, 3 Viscoid and viscolith, 166 Viscose, industrial applications of, 163, 168, 179 — sizing of paper, 163 — see also Xanthates of cellulose Wood, photographic activity of, 178 Wood-pulp, 149 Xanthates of cellulose, analysis of, 95 preparation of, 98 reversion of, 99 solubility of, 97 viscosity and stab ility of, 104 Xanthogenic acid of cellulose, 13, 93 ABERDEEN: THE UNIVERSITY PRESS