^i?S^i^H[E .3^Ai4iATIQIS[ : : OF : RAW SILK '•*. .■•-''.i !=/-■,. V. *■'■.' :■:■;■ '■ '■ »*..■;.i•<«•^?v.■■ mm Wm. Class l^^Ajb^ Book ^fi ^^, \ ^^^ ^ CDraRicEcr DEPosrr ADOT.F ROSF.X/.WF.TG SERIVALOR The Valuation of Raw Silk By Ai3ULF ROSliXZWEIG First Enci.isii Edition, 1!)1T Rc-writlen, Revised and ]uilarged to include the most recent experiences of the Author First f^iihlislu-d Si'rially in The .liiicricaii Silk jounuil Ni:w York Clifford & Lawtox, Publishers Copyright, 1917 Clifford & Lazvton A"" MAR 20I3J7 ©CLA 455977 ^^ mm . ^^ 1 M ^® ^ INTRODUCTION X ])resenting to the silk industry of America the only English translation of "Serivalor; the Valuation of Raw Silk," revised and re- w ritten b\- Adolf Rosenzweig, the author, the i)ublishers feel that they are rendering a service that will meet with the appreciation of silk men and ])articularly of the student of silk. ATr. Rosenzweig has an international reputation as an authority on silk. He has devoted the greater part of his life to research work and experimentation in determining the true valuation of raw silk. His "Serivalor" is not a theory, but a practical and logical system of standardizing raw silks, and it was estab- lished by him only after years of work along the most practical and scientific lines. 6 S E R I V A L O R A growing interest in Mr. Rosenzweig's work was aroused by the publication of the essays awarded prizes in the Prize Silk Essay Competition conducted by the Silk Association of America in 1914. In some of these essaA's Mr. Rosenzweig's work was several times quoted from and referred to. Tt was then learned that certain firms were so much interested in securing "Serivalor" for use in their own organiza- tions that they were proposing to have an luiglish translation made from the last foreign edition, long out of print and copies no longer obtainable. Believing, therefore, that Mr. Rosenzweig's work should be rendered available in English for the use of American silk manufacturers, the publishers arranged with him for this revision and in agreeing thereto ]Mr. Rosenzweig welcomed the opportunity to include in his new writing of "Serivalor" the fresh experience he had obtained in the ten years since the publication of the last edition. "Serivalor" in its new revised English form Wcis first published serially in The Atnerican Silk Journal, beginning in .that publication July. 191."), and conclud- ing in the May, ]91(i, issue. The author's final re- vision is here presented in the belief that it will not only be a valuable acquisition to every silk man's library, but that it will prove of practical value to every student interested in the standardization of raw silk.' Mr. Rosenzweig was writing this revision of his SERI VALOR 7 book at his hoiiie in Milan. Italy, during the first year of the European war and was later forced to close his Laboratory Serivalor and take up teni])orary residence in Switzerland, being a native of Austria. Laboratorio Serivalor, jNIilan. Editor Ami:ricax Silk Jourxal. Dear Sir:— In ausz^'cr to your proposition to pub- lish an English translation of my book "Scrivalor," I want to say: The book consists of a mathematical and an empirical part. Wliile the former is, of course, as valid as ever, I feel that the second part can be enlarged by the inclusion of the experience I lun'c had since the publication of my hook. I am ready therefore to re- zcrite the book for your Journal, making use of the studies and experiences I hare had during the last ten years. In my manuscript to you I shall therefore bring thezi'ork up to date and heighten itszcorth considerably. Of course, silk inspection cannot be taught by zvriting only, ajix more than can weaving or swimming, but the book should become an indispensable guide to all of those 7cho are interested in the matter. Herewith I hand you the prospectus and first chapter of " Scri- valor." Believe me, dear sir. Yours most faithfully, Adolf Rosenzwlig. **s? ^^^P k^ =*^ ^ ff^ m ^^M^ 1 m s ^^^ i/^ \< 6i 8^^^ U i ?ii PROSPECTUS WH I^ X . after ten years' studies, I was called in the year 1U02 to the managing of a broad goods manufactory and to the buying of raw silk, I tried, first of all, to find an expert whose experience would enable me to confirm ni\- thecjry. This, however, was not easy. There were some who told me of others that knew more than themselves, but when I addressed myself to them, they again pointed out others of superior experi- ence. I proposed the following trial : Between the best and the lowest grades of raw silks there is a difference in price of about $'2 ; dividing this dift'erence into forty degrees of five cents each, it follows that there must exist at least forty dift'erent qualities. Hav- ing before vou ten bales of the same color, could you 12 SERIVALOR give to each of them a distmct number of quaHty? And when, after a week, I should lay before you other samples of the same bales, would you call quality twenty again what you called it before? Or, how far would you dififer? By two, by three, by five degrees? All those to whom I addressed myself told me that this was an impossible task. Of the truth I was informed by an Italian reeler who is considered to be one of the best experts in the land of silk, who said: '"We are judging raw silk according to certain outward characteristics: purity, color and the luster of the thread, and a certain elastic resistance of the skein against the pressure of the hand. By these marks I can recognize the origin of the material, and, as I know by long experience which provinces are producing good and which low qualities of silk, I am able to judge the quality with a certain jirobability. But there is no absolute certainty. You are asking me whether my 'Extra' quality may be safely used through reeds of a certain fineness. How should I know ? I am no weaver, and all I can say is that I am having it reeled out of the best cocoons, al- ways of the same origin, and by well-trained reeling girls. And as my clients are satisfied I must suppose that I am giving them as good silk as they can expect." I asked him whether this way did not lead to occasional disappointments. He said it did. Every new crop or new crossing of breeds, brings about a new uncertainty. It is true that the thread of very bad cocoons will S E R I V A L O R 13 break during the reeling. But this is not sufficient basis for sharper distinctions. "For the final certainty we must rely on the manufacturer. If he does not com- plain, we know that the silk is good." It results, therefore, that the valuation of silk actually is at this point — that the weaver is relying on the reeler's knowledge, while the latter is relying on the weaver's judgment. In fact, if a manufacturer was offered the same material that he pays a fancy price for at the reeler's, at a much lower price by a third person, he would not dare to buy it unless it could be proved by a trade-mark that it really is the same ma- terial. Nay, the reeler himself would not be able to recognize his product with certainty if the trade-mark should be wanting. Both of them, therefore, are not judging the material itself, but are dependent on out- ward signs. It is easily to be imagined after this of what use inspectors can be who for the most part lack the experience of the reeler as well as that of the weaver. The impossibility of recognizing the quality is also an explanation of the fact that even honest reelers are unexpectedly producing bad silk, and con- tinuing to produce it until complaints from their clients make them start from their unconsciousness. If the manufacturer has no certainty about the quality of silk, neither has he of the quantity. It is of no use to him that the Conditioning Houses are giv- ing him the exact measure by which he buys, the weight, when thev are not able to give him at the same time the 14 S E R 1 \' A L O R exact relation of this measure to the one he is selhng by, the length ; or, in other words, they are not able to give him the exact size. If a bale declared to be of size I I.OO by the Conditioning House is in reality size II..")--) (and it will be shown later on that a similar error of four per cent, is (juitc common), the manu- facturer will be able to warj) only ninety-six instead of 100 warps that the real size 11. 00 would have yielded, and when the bale is consumed, he will have lost as much as if he had received only ninety-six for each mo pounds he has jKiid for. The annual loss to I he American silk manufacturers resulting from this fact ma}- l)c calculated to l)c at least a million of dollars, as will be shown later on. Another consequence of the uncertainty of size is that sometimes the pieces of goods sent out in execution of an order are eight per cent, lower in (jualily tiian the sample piece, the latter having been woven of real size 1 l..")0; for instance, the former nia\' be !;!.•')(». while the raw material for both was declared size 1 1.00 by the Conditioning House. What manufacturer has not suffered material and moral losses b\' such exjicriences ? Of all tlioe circumstances I had been well aware in r.iii-^. and 1 was sure of having found a system by wliich I could determine the real (piality and the real size of silk. In 1!>01 1 published a l)Ook entitled "Scri- \alor" and in I'JOT established in Milan a homonymous laboratory, exclusively for this task. Invited now to publish an l-Jiglish translation of my book, I want to S E R I V A L O R 15 enlarge it by the experience of these later years. In the year 15)02 I was only a weaver, though one that was making use of scientihc methods. In the mean- time I have become a reeler too. Of many facts of which I had only been aware of the effects, I have been able to recognize the causes, and the new revised edition of my book will profit by this, my enlarged experience. In the first chapter I am going to explain thor- oughly the important question of size and to show the way by which everybody may repeat the sizing of the same bale a hundred times and more without having it present, and without costs. One will see by what laws the matter is governed, and will recognize that the actual error is. in the average, of about four per cent., •while the system "Serivalor" reduces it practically to 0. Passing to f|uality, it will be shown that the differently named defects of silk may be considered as various forms of the same original defect, but that neverthe- less it is necessary to examine them, with regard to their effects on the loom, from seven different stand- points. The methods by which the Laboratory Serivalor is executing these examinations will be explained, and it will l)e shown how silk is classified with regard to each of these seven items in such a way that the best receive degree I'O, the worst H»'0. while the inler- mediate nine degrees are divided into tenths, so that in the whole we have a gradation of ninety tenths of de- crrees. After this comes the difficult task of uniting IG S E R I V A L O R the very often contradicting seven single results into one final number ; the "Resultant"' expressing the com- mercial value of the tested silk: An aim which the laboratory has been able to arrive at only after many years of trials and labor. It will be shown, further- more, how from the "Resultant" there can be derived the forty different qualities of which we spoke in the beginning, and how by this there can be calculated not only whether a bale of silk is "Double Extra," "Extra," etc., but also what price it is really worth according to the last quotations. Five per cent, of the cost price may be gained by this exact cjualification only. A still greater advantage will be derived by the manufacturer from availing himself of this principle: Every silk is good enough for the purpose corresponding to its qualities. But even less good silk will give as much satisfaction as the best, if employed in the right way, that is to say, for what it is fit. The testing being made from seven different points of view, and the silk being in this way, as it were, photographed, the buyer is able to choose the kind that possesses the qualities necessary for his purpose, with- out being obliged to pay for those that are useless to him. It will be demonstrated at the same time why the methods of testing silk heretofore used cannot possibly yield any satisfactory result, and that progress is im- possible under those methods. It will be shown how deceptive the so called "Elasticity" (in reality, due- S E R I V A L O R IT tility) is, and how little value is to be given to tensile strength. Finally, there will be given some hints which, though not belonging to the main subject, of the book, may still be useful. For instance — On the increasing of size through the shortening of the thread by throwing. Regarding the causes of the "lousiness" of silk as manifested after dyeing. About em])loying the various qualities of silk ac- cording to the fineness of the reeds. The right calculation of the cost-price of tissues, and how by this it appears that many silks that cost less by the pound are in reality more expensive than others costing more. DifYering from most books about silk that some- times repeat wrong indications from older handbooks without painstaking and expert analyses of them, this book is founded exclusively on my own experiences together with a thorough knowledge of the respective silk literature. Conscious of the fact of writing for busy readers, I shall try to be as brief as is consistent with my subject, and considerate also in treating matters for the com- prehension of which some knowledge of mathematics is necessary. Where this is lacking, however, even long explanations would be of no avail. Such parts are followed by "practical hints" for the use of the merely practical man. ClIAPTKR I THE SIZE (Lr: Titre) WE TOLD in the introduction how imi)ortant it is for the manufacturer to know the exact proportion between weight — tlie measure by which he buys — and length — the measure by which he sells. This i)ropor- tion is expressed in the size. It is established by international conventions that size is to be considered : "The weight of J50 meters (about 500 yards) expressed in half decigrams." This formula sounds rather awkward, and it contains, moreover, the germ of an error that influences the logical consequences of sizing, as will be demon- strated later on. Expressed in a clearer way, the formula is: "Size is the weight in grams of 9,000 meters (about 10.000 yards)." The object of sizing is, therefore, to indicate with S E R I V A L O R - 19 an exactitude within the narrowest hmits of variation the number of meters, or yards, contained in the bale tested. But as such a bale contains, considering only the usual sizes of 8/10 to 20/28, from ;3;3 to TOO mil- lions of meters, it is obviously impossible to measure them. On the other hand, the weight of the thread changes not only with every 4.")0 meters, as might be supposed from the official standard measure, Init with the shortest length, as the following ex])eriment in- dicates : Take a thread of about a yard's length, divide it with greatest care into two exact halves, and weigh these on a common chenn'cal balance of 1/10 mili- gram's sensibility. The tzco liolrcs zcill sliozc different zcei(jhfs. But there are chemical scales of 10 to lO'J times finer sensibility, by the aid of which one can repeat the experiment with pieces 10 to 100 times smaller, and the weight of each jjiece will always dilter. By this it ai)pears evident that the length of a great quantity of thread cannot be fixed by mere measuring and weighing, but only l)y a combination of these with mathematical methods knt)wn as the Theory of Probabilities. That our ancestors who created the -rules of sizing were not ac(iuainted with this theory is not their fault. But that they did not consult a mathematician in this difiicull niatler, they may justly be reproached for by the manufacturer. For it is he, and not the reeler, the throwster, ()r the tradesman, who has to suffer the immense loss conse- Cjuent on this uncertainty. 20 SERIVALOR It can hardly be the object of this book to teach the Theory of Probabihties but I shall try to give all the hints useful for the comprehension of the matter in hand. The studying of these elementary rules, and especially of the fourth, regarding "Constancy,"' is in- dispensable to anyone wishing to solve the problem of sizing. ELEMENTARY RULES FOR CALCULATING PROBABILITIES. Throwing uj) a coin ten times and finding that seven times it fell head and thrice tail. I evidently would commit a gross error if therefrom I should conclude that the coin will always show more than twice as many heads as tails. This error would be ihe consequence of an inadequate number of trials. First rule: Tests, whose results lay in the realm of Probabilities, must be repeated many times. Repeating the experiment 1,000 times, I get 498 heads and ."302 tails. Does this prove that the coin has a tendency to fall on the tail side? No! Of this follows the Second rule: Attempts at probabilities do not yield as clear figures as arithmetic calculations. Comparing the sums of each ten trials, I find that now the heads and now the tails are prevalent. From this I conclude that the coin will fall as many times SERIVALOR 21 on the tail as on the head side, and this result is quite as exact as that of any arithmetic calculation. Third rule: The calculation of probabilities give's as reliable result as any arithmetic ])rocedure, if these results are logically worked out. Fourth rule: The results of probabilities are to be considered as exact, if on essays repeated many times those results differ only within narrow limits from their average, that is to say, if they arc constant. The interval between the extremes indicates the num- ber of trials to be made. Till now we have worked with the object having two sides and arrived at the true after about 100 trials. Proceeding however to a die, which has six sides, we tind that 100 essays yield no constant results. ()f this follows the Fifth rule: The number of essays necessary for tindiiig that which is true increases with tlie number of possibilities. The third rule proves that a sizing sufticiently ex- act for practical purposes is possible ; the fourth that the methods hitherto used are inadequate; the first indicates the reason thereof. The fifth shows that regular and irregular threads, when treated with the same methods, will give results of dift'erent exactness. Let us now return to practical experience, and let us suj)i)ose that the owner of a bale of silk accompanies 22 S E R I V A L O R this bale to the Conditioning House, demanding that the testing should be done in his presence so that he may assure himself of its exactness. The director ex- plains to him the difficulty of the task ; the silk thread is considerably lengthened by tension and nevertheless it is necessary to stretch it, in order to measure it. It must be dried out to 0° of humidity and weighed on a very sensitive balance ; this balance must be regulated daily, as it might have changed during the night, etc. He then, draws one skein out of the bale and with the greatest possil)le care measures olt one meter. After having dried, weighed and calculated it, he de- clares the bale to be of size 0. The owner of the bale is surprised. He has bought it as 13/15. and he possesses experience enough to judge by the mere touch of the skeins that it can hardly be of size 9. But the testing has been done in his presence with care and accuracy, and he does not know yet where the error lies. He takes away his bale, and returns the next day with another one and makes the same request. Again the procedure is repeated with the same conscientious care ; result : The bale is declared to be of size 18. The owner now confesses that he has twice brought the same bale for testing and does not con- ceal his discontent. The director replies coolly that he had performed the testing according to his pre- scriptions and with the greatest possible exactness, and could do no more. S E R I \^ A L O R 33 The reader has long before recognized where the cause of the error hes ; in the insufficiency of the length measured. x\nd how has it become evident that the results are wrong? By the vast difference beti^'een them. But is he persuaded that there is no difiference between the results of the actual official sizing, that is to say, that they are "constant f Of course, the Con- ditioning Houses are not measuring one meter out of one skein, but 20 to 30 times 450 meters out of 5 to 10 skeins (the respective rules are different in different places) that is, fron: !J to l-'Vo kilometers. Why just this quantity? Why not 1 km. or 2, or 5. or 20? Why just 5 or 10 skeins, why not 2 or 15? It is generally supposed that, in fixing the rules by which the Con- ditioning Houses arc obliged to work, a quantity was adopted that could guarantee the constancy of the re- sults. P)Ut this is not the case. In the German and French editions of my book, cf 1904, I proved the unreliability of the actual meth- ods of sizing. On the occasion of the International Silk Congress of Torino, 1911, the Milan and Como Conditioning Houses published a pamphlet: "Qitcl- ques rciiiarqiics . . . siir les )noyeiuies ct Ics ccarts de titrc dcs (jreges" in which they show by the trials made on more than 30 bales, how inconstant their indica- tions are, not only in regard to size, but also in regard to the "ecart," that is. the distance between the lightest and the heaviest sizing skein, which distance generally, though wrongly, is considered as a measure of the 24 S E R I V A L O R regularity of the thread. Not only the seven Japan bales tested but also the 27 Italian bales, among which there were 13 Extra, showed vast differences, for the same bale, in regard to the size, the extremes, and the reeling. Table 7 of the pamphlet, for instance, gives the results of an Italian Extra bale, with fluctua- tions in averages of sizes between 13.70, 13.73, 13.83 on one hand, and 14.76, 15.00, and 14.00 on the other. The extremes vary between the minima of 11^^, 12, 12><, and 13, and the maxima of 15>4, 16, 16>4, 17, and 173/2. The breaks in the reeling vary between and 5, etc. Other bales show greater variations still. On table 25, for instance, (Italian ler ordre) the breaks vary between 6 and 20! On table 28 (Japan 13/15, 13^) the averages of size oscillate between 12.9 and 14.4 (that is, nearly 12 per cent.) the minima between {> and 11 3/^, the maxima between 143^4 and 19, so that a casual combination of 113^ — 143^2 (ecart 3 den.) on the one hand, and 9 — 19 {ecart 10 den.) on the other might have been possible. The pamphlet arrives at this conclusion : "1. La plus grandc partic, on Jiicinc la prcsque totalite, des soies qii'on acccpte eonune rcguUcrcs, ne Ic sont effectivemcnt pas. "2. L'essai, comme il est pratique aujourd'hui, sur un trcs petit noinbre dc fiottes, est un veritable jeu de hasard. S E R I V A L O R 25 En poisant, que dc tcllcs conclusions sont la resiiltante d'ltn grand nombre d'essais serieiis et in- contestables, nous nous demafidons, s'il est juste, et mcnie honncte, de contimier avec ces systemes." In English : "1. The greatest part, nay, it might be said, nearly all the bales of silk that are accepted as regular, are not so in fact. ''2. The testing, as it is done nowadays, on a very small number of skeins, is a mere play of hazard. "Considering that these conclusions are the result of a great number of elaborate and indisputable tests, we must ask ourselves, zchetJier it is right, or even honest, to go on with these methods.'' How can this intolerable state of things be rem- edied? In order to arrive at an answer to this ques- tion, let us make the following experiment : We divide 5 kilos of Tsatlee (Gold-Kilin) into skeins of 450 meters and receive 4236 skeins. The real size of the 5 kilos is therefore 23.(5]. Weighing every single skein, we find all sizes from IV/2 to 42. Weighing two and two together, we find all sizes from ]55^ to 351^. 26 S E R T V A L O R We see that the error has diminished on both sides, so we are on the right way. Taking 4 skeins (ISOO meters) together, we find all sizes from IS to SlVz 8 '• (3C00 •' ) " " " 19 " 29 " 16 '• (7200 " ) " " " 20!^ " 27 " 32 " (14400 " ) " " " 21 " 26 We are now arrived at about the number of skeins, by which the Conditioning Houses estabhsh the size, and we see that the real size 23.61 might, according to the casual grouping of skeins, be declared as any- thing between 21 and 26. Nay, the difference might be greater still, as in the Conditioning Houses the sizing skeins are taken from a small number of original skeins. The averages of 04 skeins ( 2S.S kilometers) vary between 22 and 25 12S •* ( r,7.6 " ) '• " 22 '/2 " 24'^ 256 " (115.2 " ) " " 23 " 2i'A 512 " (230.4 " ) " " 23J4 " 24 The following table unites the results obtained : Greatest difference Sizes from real size in obtained. % of the latter. from 13, '72 to 42 78 " 1514 " 35 51 " 18 " 311^ 34 " 19 " 29 23 " 20;4 " 27^ 16 " 21 '• 2C> 10 " 22 •• 2.5 7 " 22^ " 24H 4K " 23 " 24% 3 " 23J4 " 24 2 Length measured kilometers. Skeini 0.45 1 0.90 2 1.8 4 3.6 8 7.2 16 14.4 32 28.8 64 57.6 128 115.2 256 230.4 512 S E R I \^\ L O R 27 Comparing the first with the last cohimn, we see that the difference from the actual or true is dimin- ished by about a third by doubling the length measured. Of this there follows that in the present case it would he necessary to measure 460 kilometers (lU'^l skeins) to hring down the difference to 1.3 per cent., and !)"vO kilometers (2048 skeins) in order to he siu'e that it docs not surpass 1 per cent. While, then, taking two skeins instead of one. we have got nearer to the actual by 27 per cent., hnally, by taking 2048 instead oi 1021, we have approached it only by 0.3 per cent. — an effect that is out of i)ro- portion to the work it requires. We arrive, therefore, at a certain point, where tlie increasing of skeins ceases to be useful, and we call this the rational limit of sizing. It remains now to fix this "rational limit." in order to find out the nimiber of skeins, viz., the lenglh to be measured, necessary for exact sizing. The fol- lowing table serves for this jnirpose, and is to be used in this way : Drawing 30 skeins (of 450 meters) from a bale and choosing the two heaviest and the two lightest ones, we divide the sum of the former by that of the latter; those being, for illustration. 30 :2r) = l.-2. Look- ing for the {|Uotient on the table, we find that .■>() skeins are sufficient only when the (luotient is 1.1!). or less. We therefore increase the numl)er of skeins to GO, com])aring now, however, the four heaviest with the four lightest ones ; the proportion being f . i. 58 :4S, 28 ^ S E R I V A L O R the quotient is 1.21, and as the table indicates 1.24, we may be sure to have approached the true within -|- or — 2 per cent. TABLE FOR SIZING WITH AN EXACTNESS OF -f- or — 2 per cent. Number of skeins The result answers Number Being to be tak en f rom the purpose, if of kilo- each of the ex- the quotient is skeins. meters. treme sizes. not superior to: 30 13.5 2 1.19 60 27.0 . 4 1.24 90 40.5 6 1.28 120 54.0 8 1.33 150 67.5 10 1.37 ISO 81.0 12 1.42 200 90.0 14 1.47 By practical experience it has been ascertained, however, that in nearly all cases 200 skeins (90 km.) are wanted and that it is hardly necessary to make the preliminary trials. On the other hand there are bales for which even !)0 km. do not suffice, but these are of lower quality and therefore rarely pay the in- creased expenses of sizing. In each case we can find out by the calculation that will be shown later on, how far the true size may dififer from the established. jM'eanwhile we are i?oing to make another experi- ment. It would be desirable, no doubt, to repeat the sizing of the same bale, let us say, of Japan V/z, 13/15, a hundred times and more, and learn from the results obtained. It is true that even a hundred sizings do S E R I V A L O R 29 not give the absolute size of the whole bale, but the average of so many trials would be near enough to the actual for our purpose. But the difficulties of re- peating the sizing of the same bale a hundred times are evident. The thing would be much easier if the whole bale consisted of skeins of 450 meters ; then we should not have to measure but only to weigh. And the task would become much easier still if to each skein were attached a label indicating its weight ; then we need neither measure nor weigh but only read. And pursuing the idea, we find that in this case zee should not zvanf the skeins themselves but only the labels. Let us imagine, then, a bag filled, instead of with silk, with labels bearing the sizes contained in a picul of Japan 1^, 13/15, varying, as we know, on one single sizing bulletin from about 11 to 18. (A second bulletin of the same bale may show sizes 10-17 or 12-20.) The bag would contain about 86,000 of such labels, among which, however, the single sizes would not be represented in equal numbers, but, as we can see from any sizing bulletin, from the finest to the middle size in increasing and from the middle to the heaviest in decreasing number. Out of this bag we should then draw 30 labels and calculate their aver- age, by which we should have performed one sizing. But as we know that the bag does not contain the single sizes in equal numbers, we must not leave out- side the drawn labels lest we should alter the pro- 30 SERI VALOR ])ortion of the single sizes to each other. This would be allowed only if we contended to exhaust the bag comi^letely, that is to say to perform about 2. 800 sizings of 30 labels each. Wishing, however, to do only 100-200 sizings in order to see how one and the same bale is represented by each of them, we must be care- ful to put back each drawn label so that we might not alter the character of its contents. Let us now go a step further. Being obliged any- how to put back any drawn label, the bag need not even contain 2S00 x 30 labels, but z^'c shall arrive at the same result if it contains only one single series of 30 labels, as this remains unchanged if we put back every drawn label and mix them thoroughly. In comparison with the sizing of real silk skeins this experiment has not only the advantage of sim- plicity, cheapness, and qnickness, but is also true, as we are in knowledge of the true size of the imaginary bale, which can never be the case with a real bale. In order then to have the reflex of reality, we need only to write the 30 figures of some sizing bulle- tin on handy labels, put these into a little bag and begin with the drawing. The bulletin No. 5136. of April 15. 1-915. of the Stagionatura Anonima, Milan, reproduced here, might serve as an example : S E R I \^\ L O R 31 Report on test made on samples of 10 skeins in raw R C 24 on weight of Kilo 0.705 SIZE. IVA 133^ 15^ 12 14 16 12H 14 16 12^ 14 16^ 121^ 14 16^ 13 141^ 17 13 uy2 17 ]3 uy. 17^ li'A 15 171^ ^VA 1 15 K- 1714 We see that this bulletin contains: Sizes n><, 12, 12K'. 13, 13^, 14, 14^^, 11 3 3 3 4 3 15, 15^>, 16, 16J/, 17, 17^. 12 2 2 2 3 Sum: 4375. average 4375:30=14.59. (Any other sizing bulletin that shows a similar average (about 14.55 to 14. (Jo) with a similar ecart (G den.) would serve as well.) The experiment, viz., the drawing of oO numbers should be repeated at least 100 times, writing down each drawn number, but calculating the average not of each 30 but of each 10 numbers, and drawing to- gether 3 averages of 10 to one of 30. This procedure 32 SERI VALOR is more practical, as later on we shall want the aver- ages of 20 X 10 numbers, and so those of 10 will come handy. According to the Theory of Probabilities among 100 sizings (of 30 numbers each) there will appear: averages up to about — 14.0, 14.1, l-i.2, 14.3, 14.4, 14.5, 3 4 10 10 10 13 14.6, 14.7, 14.8, 14.9, 15.0, 15.1. 10 3 7 7 13 10 Of this follows : 1. Of 100 sizings of a bale of real size 14. G to 14.7, 50 will ])c such that the bale may be delivered as 13/15. according to "usage." 2. If therefore I have sizings made on 100 bales 14/15, or 14/lG, I may be sure that 50 of them may be delivered as 13/15. Repeating the sizing of the remaining 50 bales, half of these will again appear as 13/15, and I have only 25 bales left, with which I go on in the same wa)^ until all the 100 bales have passed as 13/15. How many sizings were necessary for this purpose? 100-1-50 + 25 + 12 or 13 + 6 + 3 + 1 or 2 = 200, that is to say two sizings per bale instead of one. 3. A more practical and more direct way is to S E R I V A L O R 33 luive each bale sized twice at once. Thus we receive two bulletins for each bale, of which nearly always one will be of 13/15. This tzvofold sizing is tlie gen- eral useage in nearly all silk-trading places. It is therefore a mathematical fact that by the two- fold sizing of each bale the real size appears altered by about 4 per cent. How large the loss accruing from this fact is, especially for buyers of Japan 13/15. was shown to me by the sizing of 112 piculs that the Labo- ratory Serivalor had to effect some time ago. These 112 piculs of Japan V/y had all passed as 13/15. Ac- cording to the sizing of the Laboratory, on a basis of 90 kilometers (200 skeins), 58 of them appeared as 11/15, and the average of the li'hole 112 z^'as 14.58, therefore more than 4 per cent, too heavy. I am sure that the controlling of each great lot of Japan iy2. 13/15, would give the same and worse results. The reason lies in this : Tn each country the great bulk of cocoons consists of a thread (baz'a) of a certain size which then determines the size of the silk-thread reeled therefrom. In Jajjan the size of the bava of the main part of the crop evidently is near to 2.45, and therefore to the reeler 4 cocoons will yield 4 x 2.45 = 9.80, or 9/11 5 -) X 2.45= 1 :.'.:.'-) " 11/13 — 12/1.3 6 " " " 6 X 2.45 = 14.70 " 14/15 — 14/16, but it is difificult for him to arrive at size 14, for which he would be obliged to employ : 34 S E R I V A L O R or 5 cocoons of 2.80 ^ 1 1.0 or G cocoons of 2.35 = 1-1.1, which kinds evidently are comparatively rare in Japan. In Italy a bava of 2.8 den. is quite common, and therefore in Italian silk the size 13/15 will generally be right. On the other hand, a great part of Italian 11/13 are in reality 12/13 to 12/14, as a bava of 2.4, of which 5 cocoons would yield size 12. is rather rare, and the recler must be content to find 2.5 to 2.G yielding sizes 12.5 to 13. (It would be easy for the reeler to find 3.0, of which 1 would yield size 12; but the reel- ing of 4 cocoons is very ditticult, as we shall see later on, and the thread contains very many fine ends.) It might be observed here that though every con- scientious reeler makes frequent size tests such tests are of no great avail, as the variations in drying the wet thread bring results more or less incomplete, ac- cording to the humidity of the surrounding air, and the reeler, of course, does not use a conditioning appa- ratus. This leaves the reeler uncertain of about 4 per cent, of the size. How great is the loss to the American manufac- turer resulting from these circumstances? The Ameri- can consumption of silk amounts on an average to 100 millions of dollars a year. We have seen that the twofold sizing of each bale changes the true size as much as 4 per cent. Supposing that this applied to only one out of four cases, which is certainly below S E R 1 \' A L O R 35 the mark, yet even then we estimate a loss of a milHon of dollars yearly. (This loss riinmng through im- perceptible holes accounts for lucniy a)i uiuiccountable irtiitus ill the manufacturer's annual balance.) How can this loss be avoided ? Simply by exact ■^izin.i;-. Taking- as a basis, instead of the nsual 1.'>.5 kilometers '(;5U skeins). !)() kilometers (200 skeins) as the Lal)oratory Serivalor is doing, the model bale IM-evionsly utilized for our experiments, will yield in each 100 sizings : Siziiigs 7 2n .-36 17 20 Sizes 14.:i.")-14.4l> iii)to14.o 14.6 14.7 14. S (These results may be controlled by any reader. We had previously said that the averages of each 10 iuiml)ers should be noted. Writing down these aver- ages on new labels, drawing them in the manner previ- ously explained and taking- the averages of each 20, these represent the averages of 20 X 10 = 200 sizing skeins = DO kilometers, viz.. the quantity measured by theL. S.) We see now how the possible error has dwindled. ( )nly : of 100 bales 1 J/l-"), and none of 14/16 could be delivered as llVl'"), and this small possible gain will induce nobody to have each bale tested twice. But also the deviation of 1^^ per cent. (1 1.35-14.40 instead of 14.59) will occur as often above as below the mark, and therefore be practically reduced to nothing. 36 SERI VALOR How great the error is in each single case may be calculated mathematically out of each series of 10 siz- ings of 9 km. (^20 skeins) as done for each single bale by the Serivalor system. The formula, which un- fortunately cannot be explained at length here, is the following: a being the average of n sizings, d the dif- ference of each sizing from the average ; then we have: M 100 /^ d 4- d ' -f d ' i-d" •2 3 n R = The resultant : R expresses the mean division in per cent, of the size. Here are some actual exam])les : Italian 1st order, 9/11. 10.0, 10.2, 10.2. 10.2. 10.2, 10.3, 10.5. lO.G, 11.2. .\verage 10.4. Mean deviation 0.33 per cent. Italian 2d order, 11/13. 11.0, 11.1, 11.2, 11.4, 11.6, 11.8, 12.2, 12.3, 12.6, 12.8. AveraL;e 11.8. Mean deviation 1.75 per cent. Italian Classical 14/16. 14.2, 14.4, 14.6, 14.7, 14.S, 1,5.0, 15.2, 15.3, 15.4. Average 14.8. Mean deviation 0.30 per cent. Italian Extra 27/29. 26.5, 27.3, 27.5, 27.5, 27.5, 27.8, 28.0, 28.2, 28.7. Average 27.7. Mean deviation 0.78 per cent. Japan Double Extra 12/13. 12.0, 12.0, 12.1, 12.2, 12.4, 12.6, 12.8, 12.9, 13.0, 13.0. Average 12.5. Mean deviation 1.3 per cent. SERIVALOR 37 Japan V/., so called 13/15. 13/7, 13/8, 14.0, 14.1, 14.2, 14.3, 14..5, 14.8, 15.6, 16.0. Average 14.5. Mean deviation 1.7 per cent. Japan V/>-2, so called 13/16. 13.7, 14.0, 14.3. 14.7, 14.7, 14. S, 15.5, 15.6, 17.0. Average 14.9. Mean deviation 2.0 per cent. Some of these examples show that for very ir- regular threads even sizing on a hasis of 90 km. is liardly sufificient. In such cases, in which the mean deviation exceeds 1 per cent., the sizing should be made on 180 km. In any case the result can be exact only if the sizing skeins are taken from a sufficiently great num- ber of original skeins. Of this number which, of course, has also its rational limit, we want to say : A l)ale of silk is j^roduced either by many reeling girls in a short time, or by a few of them in a longer time. In the first case the number of skeins drawn ought to be equal to the number of reeling girls, in the second to their number multiplied by the number of weeks they had worked at the bale — (supposing that in the worst case their way of working had changed in the course of a week). For according to the laws of probabilities the re- sult of any new test will be equal to those of the former ones, if the skeins drawn represent the entire char- acter of the bale. yVnd as the latter is the product of a certain number of individuals, it would be de- 38 S E R I V A L O R sirable that a skein should be drawn for every indi- vidual. It is necessary therefore to draw the skeins from all parts and layers of the bale. Of course we do not know which is which ; but as they are indis- criminately mixed, by drawing them from all parts, we may be sure to get a true characteristic of the whole. Let us now consider opposite suppositions : a. The bale (100 kilos) was reeled by one single girl. With a production of 500 grams daily, the work should rccjuire oO weeks. Therefore oO skeins would be drawn. b. The bale was finished in a week. Then 30 girls have worked at it. and we have again to draw 30 skeins. For these reasons I made the tests first on 30 skeins, and having arrived at satisfactor)- results, tried to diminish their number to 20. This number has proved sufficient as far as Italian silks are concerned, but not quite as reliable for Asiatic silks. The Laljoratory Serivalor therefore i)roceeds as follows : Of eacli bale (100 kilos) Italian, 20 skeins are drawn. " Picul (GO " ) Japan, 15 (^^^e have said that of Italian silks the skeins must be taken from all parts and layers of the bale. SERI VALOR 39 With Japans each skein must be drawn from a dif- ferent parcel. In testing Japan silks, two piculs arc always coupled together and the testing is done on 2 X 15 = 30 skeins. These piculs must be considered as a unit and employed together on the loom. Of each of the 20 (respectively 30) skeins 4,500 meters (^10 sizing skeins) arc measured off and the size established according to their weight at TO per cent, humidity of the air. The silk measured off this way then serves for all the other tests to be mentioned later. (Anyone who may doubt of the measuring being really done may demand that an additional 90 kilo- meters should be measured off and sent to him.) Chapter II SOME REMARKS ABOUT QUALITY IN GENERA L IF A dozen experts were called, let us say, by a judge in a lawsuit, to answer the question, "What does the quality of silk consist in?" they would in all probability give twelve complicated answers rich with technical expressions, each absolutely ditierent from the other. Dark hints about "a certain touch," "a certain luster," etc.. would only serve to hide the want of a clear definition, and after having heard the twelve, the judge would be as wise as before. In reality, the answer is very simple : The best silk is the one that alloics the highest speed on the loom. On the loom, and not in winding, warping, or throwing, for these procedures are not so important as weaving. If a bale of silk proved bad in winding, warping or S E R I V A L O R 41 throwing, but allowed speedy weaving, it is of excellent quality; just as it is of bad quality if the weaving is slow, though the winding, warping or throwing had been excellent. This, of course, appears evident, as a break in winding and throwing stops one thread only, in warp- ing 300, in weaving up to 15,000. In fact, the wages and general ex])enses are in the proportion of 1 for winding to 'ijA for warping, and 10 for weaving. The speed in weaving is worth ten times that in winding, and four times that in warping. But there are many people who consider the wind- ing as a touchstone for quality in general. To those I want to say that this supposition is quite wrong, as will be demonstrated in its place. It is an erroneous supposition as well, that the better silk will yield a better, a more durable tissue. The durability of the tissue has nothing to do with the quality of the silk it is made of, and "Gold-Kilin," for instance, will yield a stufT that will wear at least as well as one made of Italian Extra that costs about 60 per cent, more, the only difference lying /;/ the ci'cn- ness of the surface. But now there arises a new question. ( )n what depends the possibility of speedy weaving? The answer is: On the "iinifoniiity" of the silk thread. The silk thread ought to be '"uniform." The ideal claim for it is that it should have the same diameter 42 S E R I V A L O R throughout its whole length. Its "quality" is pro- portional to its "uniformity," that is to say, the more uniform it is, the better is its quality, and vice versa; and all its ditTerently named defects are only various forms of one and the same defect : want of uiiifurmity. But according to the shape in which this defect appears, its influence is ditYerent on the speed of pro- duction, and therefore it will be necessary to treat of these various shapes in separate chapters. In this instance I want only to give some hints about the way in which the reeler must try to arrive at the greatest possible uniformity. The silk thread cannot be reeled out of one co- coon, but of four at least, better from five of them united. Here follows the application of the first rule: Tiic iiiiiiiher of cocoons nuist be the same during the -u'hole time of the reeling. Within a lot of the same race the diameter of the cocoon's thread is about proportional to the size of the cocoons. Hence the second rule : The cocoons reeled together ought to be of the same race and of about the same sice. The diameter of the cocoon thread is very variable, but as a rule it is much thinner at the beginning and at the end than in the middle of each cocoon. This fur- nishes the 3rd Rule: llie cocoon threads must complete each other in sucli a manner that about Jialf of their nuin- S E R I V A L O R 43 bcr arc at the beginning ur at the end, u'hile the other half are at the middle. If several threads are to be united to one of uni- form diameter, it is indispensable that each of them should hrst form a strais,dit line. On the cocoon, how- ever, the thread is laid in circles and looi)s like the fig- ure th Rule: 71ie serici)i inust not be soaked too much. The reeler. however, can act according to rules 5 and (J only when every lot of cocoons is assorted in such a manner that only cocoons of the same texture are emploxed together. I'or the less tightly the worm has crossed the thread, tlie deeper the hot water pene- trates, and the more soaked will be the sericin. The chief differences of texture arc called "Reali," "Rea- 44 S E R I V A L O R lini," "Scarti," "Bombaggiati," but there are still num- erous subdivisions. The 7th rule is, therefore: The cocoons must be assorted according to their texture. The assorting, however, is not sufficient for its purpose if the speed of the reeling is not regulated according to the (juality of cocoons. llie right speed of reeling, then, forms the Sth rule. The well stretched thread laying itself alongside its neighbors, does not unite with them into a round cord, but forms only a flat ribbon. Hence the 9th Rule: The stretched cocoon thread must be united by a concentric pressure. The now completed thread is wet and much in- clined to sticking, adhering with its neighbors, espe- cially where they rest on the six logs of the reel. The 10th rule therefore is : The sticking together of the threads must be avoided. To sum up what has been said before: The more thoroughly a rceler acts according to the ten rules established above, the better zinll be the quality of silk produced. We are now going to consider the consequences of each single form of the original defect in silk, and to show that they are not to be recognized by the meth- ods used hitherto ; but that there are other methods that will vield reliable results. Ciiapti:r Til R E G U L A R 1 1" Y UXDKR this title we are going to treat that want of tmiforniity which becomes a])parent in the varia- tions of size. -\s we said in tiie preceding chapter, these variations are the consequence partly of the irregularity of the cocoon thread, and partly of its being reeled without observation of the first and second of the rules established there. This irregularity is considerable, and more so with the thread of larger cocoons than with that of smaller ones. The difterence within one cocoon niav be ex- pressed by the proportion 1 :-'! ; that is to say, cocoons of the average size of 2 :S den., for instance, contain all sizes from lA to 1.'3 den., the fmest occurring at the end of the thread (whose length is about 500 meters) the thickest in the middle, the middle sizes at the beginning. From the proportion 1 :'■) and from the mathe- matical law of "Combination" it follows that tlic unit- 46 S E R I \^\ L O R ing of five or six cocoons must yield a thread whose variations of size cannot be of smaller proportion than 10 :]."), and that consequently even an excellently reeled J 3/1. "3 must contain all sizes from about 11 to K. In fact, even in a well-reeled skein of lo/\o, of about 50 yards length, we shall find : of size 11 12 13 14 15 10 17 about 3% -10% 20% 40% 15% lO^/c 2% This table expresses the smallest theoretical pro- portion, which, however, is always exceeded in skeins longer than 50 yards. By this we can see how wrong the trade usage is: that a bale of 1;?/1.j allows no greater ecart than •") den. ; when in reality it always exceeds ('> den. This error results from the habit of finding the ecart by skeins of loO meters, which ecart is not only wrong but highly inconstant, as proven bv our experiment of I0() sizings of 30 labels, and which is also to be seen on every page of the publication of the ]\lilan and Como Conditioning Houses, which we referred to in the first chapter. In order to find out the length that should serve as a basis for calculating the variations of size, it is necessary to look at the work of reeling. The reeling-girl is spinning I to (i threads at a time out of 30 cocoons, with a si:)eed varying, accord- ing to the ([uality of cocoons, from 100 to 130 meters in a minute. This speed is so regulated that the girl S E R I V A L O R 47 has time to "'cast" a new thread at every break, which breaks average lo a minute. In the time between one "cast" and the other each of the six threads runs there- fore about l"i<» meters, and this length represents 15 the average of the defect caused by the reehng-girk If, then, we want to test the thread with regard to this defect, we must, according to theory, divide it into lengths of 5-10, but not 450 meters; this latter length would serve only if the reeling-girls were working 50-100 times slower than they do, that is to say, if they were nearly sleeping. In fact, the testing must be done by lengths not over 50 and sometimes down to 5 meters. Even under this condition it must never be based on the extremes, as these are never "constant" even when established by a very great number of essays. There are even mathematicians who do not consider the extremes at all. After many practical experiments the Laboratory Serivalor, Iiowever, has come to the conclusion that, for our purpose, the extremes are to be taken into con- sideration as well as the other elements of testing for regularity. Continuing now our research for the smallest pos- sible variations of size within the thread, we find that they are limited toward the thick side. A reeling-girl that has an order to reel five cocoons has no honest reason for taking six of them, and the thread therefore 48 S E R I V A L O R ought to show no greater thickening of the desired size than is justified by the preceding formula of 10 :15 ; nay, the formula must in this case be reduced to 13 :15 as the thickening of the cocoon thread is less important than its thinning towards the end. The formula 13 :15 is to be understood in this way, that for the average size of 13 the heaviest size ought not to be over 15, in other words : tJie thickest parts of a thread ought not to surpass its average size by more than 15 per cent. This theoretical claim is justified by the fact that threads of this regularity occur, but they are uncom- monly rare, which proves that either the cocoons are, in general, not assorted with sutBcient care, or that the controlling of the reeling-girls is not as strict as it. should be. For the girl will take six cocoons instead of five whenever for a certain time she has allowed the reeling to go on with three instead of five cocoons. As the controlling of size is done by skeins (provino) of 450 meters, she wants to bring about the right size of the latter by adding to the size as much as she has lost by her carelessness. This ought to be strictly avoided, but it has become so common nevertheless that it has brought al)out the adage : Tra il grosso cd il fiiio Sorte il provino. ("What with a thick thread, what with a thin one, springs forth the provino.") Toward the thin side the possible deviation is S E R I V A L O R 49 greater, for to the natural irregularity of 10:13, or 77 :100, are added the unavoidable breaks of the co- coon-thread, which, as we have seen, take away at least 1/5, sometimes 2/5 of the size, as long as they are not mended. (In reeling four cocoons this frequently occurring accident diminishes the size by 2/4^50 per cent., and therefore at least five cocoons should be em- ployed, as mentioned in Chapter I.) In fact, the variations toward the thin side are in the best case not inferior to 30 per cent., which corre- sponds to the coincidence that the reeling-girl missed a "cast," while the thread was running i)0 per cent, of its regular size. We have seen, then, that the testing for regularity can be done only by lengths of 5 to 50 meters at the highest. Now it remains to fix the number of these pieces necessary for reliable testing. This number can be found only by experiments. By many of these it has been ascertained that, as a rule, at least 200 and for very irregular threads at least 400 are necessary to obtain constant results. The calculations to be made with this material can- not be explained here, being intelligible only to mathe- maticians who know ihcni as the "Theory of the least squares." To any of those exj)erts who may read this I want to say that the greater or smaller difference be- tween the results of the arithmetical and the geomet- rical series of differences gives a reliable indication. 50 S E R I V A L O R whether we are in presence of a bale of "Natives'' or of "Filature." With the latter the difference is always below 5 per cent., with the former above 5 per cent. Though a mathematical explanation is impossible in this place, the procedure can be exhibited to the reader by graphic reproductions. Those given here are made for size 13/15 on skeins of 20 yards, and are applicable for this size only. For the conditions of the regularity of the silk thread are such that the thinner sizes can be made neither as regular nor as irregular as the thicker ones. This will appear evident to anybody who reflects upon the irregularities that must occur, and those that may occur, with four co- coons on the one hand, and eight cocoons on the other. Conse(|uently the distances between the degrees 1-10 of the gradation Seri valor are different for different sizes. The working girl charged with weighing the little skeins of 20 yards is registering them on a sheet of paper on whose left hand margin the sizes 4 to 33 are printed in a vertical row, while to each of these figures corresponds a horizontal row of numbers from 1 to 60 (see plates). The first skein being for instance of size 14, it is registered by a dash through number one of series 14, the second of size 17, by one through number one of series 17, and so on, until all the 200 skeins are registered. Tt is evident, that the more regular a bale of 13/15 S E R I V A L O R 51 is, the more frequently will occur the size 14, and after it 13 and 15, while, the more irregular it is, the less frequently these three sizes will turn up, while those distant from the average will increase proportionately. Having finished the registering, the dashes will form a triangle, whose basis will be the shorter and whose height the greater, the more regular the bale is, while on the contrary the height will be the lower and the basis the larger, the more irregularly it was reeled. (For all these, and all the following conclusions it is an essential condition that the silk serving for the test should be taken from 20-30 skeins, drawn from all parts of the bale. Carelessly drawn skeins, or a smaller number of them cannot give reliable results.) By com])aring. then, the basis to the height, we arrive at a gradation of regularity, and even the aspect of the diagram suffices for giving a general idea of it. This comparison, however, is not sufficient for a grada- tion by tenths of degrees, as the Laboratory Serivalor is calculating it. Let us now look at the following diagram, repre- senting "Degree Serivalor" (S") 1 of regularitv. for size ];5/15. Di,\(;k.\M 1. We see that the -^00 skeins are divided as follows : Size. . . . . . 10 11 12 l.S 14 15 1() 17 .Skeins. . o s 30 :is id 46 29 1 52 SERIVALOR Diagram 2. The diagram, rei^resenting S" 2, contains : Size 9 10 n 12 13 14 15 16 17 IS Skeins 2 4 10 29 34 43 38 29 10 1 We see that the number of the "ideaF' size 14 has diminished from 46 to 43, while the basis of the triangle is enlarged by the sizes 9 and 18. But we see also that size IS appears only once in 200 skeins, which proves that a smaller number of them could have been suffi- cient only in exceptional cases. The following diagram represents S'* 3 : Diagram 3. It contains : Size S 9 10 11 12 13 14 l.'j 16 17 IS 19 Skeins.... 2 4 4 S 28 34 40 36 25 11 7 1 It is to be observed how the gliding down of the center sizes is increasing, while the extremes are ap- I)earing only in small numbers, a proof that no judg- ment can be based on these alone. Compared with the preceding diagrams, f. i., they are but slightly dif- ferent ; in size 18 the difference is already more sensible (7 to 1) and so on for the other sizes, which demon- strates that the proportion of all sizes must be taken into consideration. Diagram 4. On this diagram, representing S'^ 4, there appears : S E R I V A L O R 53 Size S 9- 10 11 12 1:5 li 15 16 17 18 19 20 Skeins. .2 4 S 10 22 30 3S 31 21 15 9 3 1 The greater irregularity is expressed by no new- extreme on the thin side, but only by the continual "gliding down" of the central sizes: while on the thick side there appears a new extreme. This is a character- istic of those silks that are nearly too irregular for single weaving. Size 8 consisting only of ;> cocoon threads, even a careless reeling-girl will nol go on this way for a long time, and consequently the in- creasing irregularity expresses itself chiefly towards the heavy side. The following diagram, representing S" 5. Diagram 5 contains : Size 7 S 9 10 11 12 13 14 15 Ifi 17 IS 19 20 21 Skeins.. 2 2 (i S 12 22 30 3(1 2S 21 15 9 5 3 1 Incessantly the height is flattening towards both sides and the center size II has diminished from li'» ( S" 1 ) to ;>(;. l)l.\(iKA.M <). This diagram, representing S" (>. contains: Size....? s 9 10 II 12 13 14 15 1(1 17 IS 19 20 21 22 Skein?.. 2 4 6 10 14 20 27 34 24 20 IG 10 G 4 2 1 The sizes 10 and 1 ], forming the extremes of S" 1, and represented there only by ( a ■* a s B CO CD o CD O CO CO s a s 8 s a 8 o 00 lA o «0 s s a 8 a a s S3 $3 c? a o CO o R S a a s a a a s a o CO o ?! a s? a a a s a o CO o n a a a a a a a o o CO s s a a a a a s a o o a s a a K a a a a a < / o i < c c 1: i ! i S s s i < / ? ? f 1 ! S i ? ? i < o o z a :; s s a a a i| a a » ■) 1 < ) 1 o a a a K a a a a a r- o •* o Z c < It xc ■A o «0 60 SERI VALOR CO < s a a s g o 00 o a s a s B a a s 2 ■» o to in o CD S a a s s a s 8 o a 8 a ?3 R S3 S 8 o lA s a s s s a s a o 00 i a K a a a a o •» 1^ o ?? a a a a a a a 00 O o OS S s a a K a a s a s a o ? f ? ? ? f f ? ? c 3 ? ? 3 3 3 3 3 o a 8 a a a a 3 a s a 1 { / 8 ? ? ? c c < c 2 ® < o i ; ; i J i ! 1 i i i < c ! ! 1 i ! ° ■ c 1 c c < e o CO 8 i o to CO a 8 a a R a a a a :^ ( 1 K 3 3 3 3 \ o a 8 s a s a a ;; a o « ^, ^ *4' CO Ov C » - C» 1 f«i i lA » * lO ( > < 1 = o O CO » -a- ■n fo ^ r s i ri J o i ? s ? 8 i S ? oo 8 \ ! Si i i < o S s S ; c < o s s s a s 8 \ < o R S a a a 8 p i S o s s a a K a s R ?: 8 o 00 s a s K 8 a a 8 o r- •♦ r^ CO 2 X *' lA •«' M «» !>• « o« 8 M S !^ a a a a a a a a a a a a a a a s S s s s s s s s /s I a s a a a a a a cr, n M. s n n s n " ; 1 " s n n n ?; n n ^D s 9, a a a a a a i la a a a a a a a a a a a a a a a i la a a a a a a a s a a a a a a a ! la a a a a a a a K s R s K K B s E r s B K B K B s 8 » a a a a a a i J a a a a a a a <3 S3 83 a a a a a a i 1 a a a a a a « S S s s a a s a a ; a a a a a a a a n a a n a a a a. s i i \ a a a a a a a a a a a a a a a ? i i i V^ a a a a a n s s s s s s s s j^ ; : ; ^ ?: s s s ^ s a a a a a a a a i ; i \ a a a a a a £ * 2 2 2 2 2 2 i ; ; : S ' £ 2 2 2 £ CO CD 2 2 2 2 2 2 1' s ! : t 2 £ 2 2 S z Z to Z ;o Z z Ij : : ' \ z z 2 2 2 •a •s. 2 2 2 2 1 : ■ - 2 2 2 £ o O, c r, c n }. , < , L2 ^ r> « 2 a S 2 2 2 =/ '. ; * ; t s ^ £! s S o o o o o o < = i ; ; c ] 1 o 2 e o a> o> o> o> !J. CT. / 1 ■ a . < \ a> o> CT o. eo a 00 CO 00 7 c < a 1 '. V" 00 OO to f* t- r- r- r- - r - ' Y t- t- f- u> f (M * 1 * • 1 <0 ^ d > « ^ o 4 Si S E R 1 V A L O R G5 ()(] S E R I V A L O R * 06 < ^ ■n s -^ ?3 S OS 00 m SI B s f CO s s < ( 1 c ? ? s ? ? ? c ? ? 3 3 3 3 3 / C S ? g f ? > ? ? 3 3 3 3 / ? ? ? ? ? ? ? 3 3 3 3 3 ^ 9 ? < * 2 ? 3 3 3 f ? c 1 s? u S I ? 1 ° \ C g < s S ? g g < ? ? 3 3 3 3 \ a ? ? $ ? g g g g V > ? ? 3 3 3 3 V g s g c V C ? ? ? g g g g g C •v e > 4 ! i 3 a 3 3 3 a g3 -a- < P4 C4 s? 9. B g3 -3- 04 S3 g3 gs ^4 ?3 s g3 <* w r- CO e ^ < » : c < l» 1 c M 1 W » 1 ^ : a ) < c > 4 r4 (M r4 r>4 04 SERI VALOR 67 ! s ?3 < ^ ^ Z O Z < 2 Q = oo s gs s a $3 o <0 s s s . ?3 a o o 7 ^ ? ? < 5 3 CO S s s o o ( 1 s s o y c R s c K ?3 CM s u ? ? ; ? 3 >4- ? ! 3 3 3 S 3 r s c ? g % % ? ? \ > 3 ? 3 3 3 3 3 ^ ? 1 3 § 3 3 3 \ « P g g < ( > e g g ? ? 5 ? 5 3 5 3 ! 3 a g: S s a a : s o i cs «M ?? a a a a a g! o r- i M s 8 a a a s a a o i CM S a a K a a a CD CM a a a a a a r- 1 \ S « ■A 10 h - CO cv ( 1 f* t 1«' ■» 9 > » c a > e ^ c 1 1 o cs M >* CM CM G8 S E R I V A L O R a o a ^ I < 2 ^ Z o ■A s s a s o CO <0 a s s K 8 a s s o CO R S S s K 8 a s 8 *" o g ? c > i o s s ? ? g s s c ! ! i i c s f g ? E c y s 8 a s 8 s i < ■ / S 8 a a a a / 1 8 8 a a a a I la 1^ ? S S ? i« ? ? = > ? ? 3 3 3 3 3 ?: \ < s ft, ? ?: S e S e ? s ? ? f f s ^ ? ? 3 3 3 3 S 3 3 s C ^ ? 9 g ? P S c s s s 8 a a 8 a s a s a p- o o CM a s a a s a a a a a a c* a R a a s a a a a a a o -3- a s a a a a s a a a o •s in «M a S a a s 8 a s a a 8 o 19 a s a a a a s a a a o / . 1 ^ r* ■ : S '■' ^ 1 1. f '^ s ♦ < in (0 f « a ) < N < » £ > c * «n XO <* « » >« « 1 c s e > 4 1 r 4 4 1 4 o IS in CM C4 DIAGRAM NATIVE. ♦x 2 3 4 5 6 7 8 9 to t1 12 13 14 15 16= 17 18 19 20 21 22 23 24 25 26 | 'V ' .^ I^^AvC^ 4__ 5^ 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 5 6x ^ 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 6 7y -i ■4. 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 7 ftk > 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 8 Q< , 18 19 20 21 22 23 24 25 25 27 28 29 30 31 32 33 9 10« „ ,„ ,. 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 10 ' 11y 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 11 12^ 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 12 ^ IS/ 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 13 14o ^ ^^ 32 33 14 "^ ^^^ ^ , - „ „ .. .. > IB 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 15 " J 1 K r> . „ „ ^r 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 16 ' ' Jr ' 1 7o 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 17 1 A^ 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 18 1 Qo 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 19 >**^' 20° ^'s 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 20 > 21 = -3> / < 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 21 22 V 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 22 23° ■4 3 4 5 6 7 8 9 10 It 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 23 2*0 '2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 24 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 25 26°! 2 3 4 5 6 7 8 9 to II 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 26 27 t 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 27 28 ' 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 28 29 t 2 3 4 5 6 7 8 9 to 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 29 30 t 2 3 4 5 e 7 3 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 30 31o| ^ 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 31 32 1 2 3 4 5 6 7 8 8 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 32 33 2 3 4 5 6 7 8 9 10 11 12 13 14 15 15 16 16 17 18 17 18 19 19 20 21 20 21 22 23 22 23 24 24 25 25 26 26 27 28 29 30 31 32 33 33 34° h 3 4 5 6 7 8 9 10 11 12 13 14 \K^ 4 b5 1 2 3 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 5 ^6''{!2 3 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 6 37 1 2 3 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 7 380I2 3 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 8 i9 1 2 3 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 9 M0° h 3 ' 6 7 8 9 10 11 12 13 14 15 16 17,18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 10 Chapter IV FINE THREADS IN THE foregoing chapter we have considered regularity chiefly with regard to the surface of the tissue, but regularity also includes fine threads, as de- fects, as they form a great impediment to speedy production and therefore require special examination. Let us call in again the twelve experts of the second chapter and ask them : At which point a thread may be called a "fine thread ?" Is size 10 a fine thread ? Or size 9 ? Or 6 ? This time the twelve answers will be more uniform, if not more satisfactory: "We do not know." In fact, the answer to this question is not quite easy, and in order to find it, it is necessary to contem- plate the work as well of the reeler as of the weaver. Let us begin w'ith the latter. The loom evidently exercises nearly the same tension "T" on the thread, SERIVALOR 71 whether it is working with size 9/11 or 29/31. It re- mains in both cases the same instrument with nearly identical qualities, and of this follows that "T" repre- sents an absolute and not a relative value, that is to say, that 'the loom requires a certain minimum re- sistance of the thread, without regard to its size. This tension can be measured by inserting a sensitive dyna- mometer into a well-mounted warp ; with results at about 25 grams for the weaving-loom, and near to .'iS grams for the lace-loom. The sizes 6.75 on the one hand, and 10 on the other, are able to resist this tension — as will be demonstrated in the chapter on Tensile Strength — and so we might establish, that for the weaving-loom the sizes below G.To, for the lace-loom those below 10 are to be considered as "fine threads." This is in accordance with practical experience. The thinnest size generally used for single weaving is 11/13 (of which we know, however, that very often it is 12/13 to 12/14) and in this size it requires no extraordinary skill of the reeling-girl not to go below 6 3/4. There are reelers, however, who produce 10/12 and even 9/11 fit for single weaving; but it has not been possible till now to produce 8/10 for this purpose. In fact, the frequent thinning of 30 per cent, that was demonstrated as being inevitable in the last chapter, amounts to 3.7 den. for size 9; consequently every bale of 8/10 must needs contain a great many passages of size (9-2.7) = 6.3, which will l)reak on the loom. While for 9/11 the size of inevitnl)le fine threads will he 10-3 = 7. 72 SERI VALOR and therefore it is possible to produce 9/11 fit for single weaving. The same way for size 14/15, that is generally em- ployed for Malincs, the inevitable thinning arrives at size 10.15, which is able to resist to the tension of 35 grammes recpiired by the lace-loom. But a conscientious testing establishment cannot top at these figures and consider its task accomplished by them, nor should it declare that a bale of 38/30 contains no fine threads, because the thimiest sizes oc- curring in it are above G 3/4. It was necessary, therefore, to establish relative measures, viz., such in proportion with the size of the tested bale, without regard to the fact that the loom is heedless of this size. Theoretical researches, as used in the last chapter, could give but a general idea about the variations of fine threads between good and bad silk ; these could be found only by practical experience. It was necessary to establish, by the work of many years and by experi- ments on thousands of bales, the proportions between the sizes reeled of 4, 5, 6, etc., cocoons, and only after having ascertained by long experience that f. i., not one bale of 19/21 appeared that did not contain at least size 14 ; this size 14 could be established as the mini- mum for S° 1, for size 19/21. On the other hand we had found that the worst bales of 19/21 went down as far as size 5, but also this fact had to be confirmed rei)eatedlv, until we could f\x the S° 10 at size 5. S E R I V A L O R Td It was demonstrated in the last chapter how the fine threads are to be found, but we said also how cau- tious we must be in drawing conclusions, having to do with extremes. Therefore we must not judge by the latter alone but must take into consideration the whole character of the triangle, and go on with our researches if the extreme found is not in accordance with the former. Also in this case experience is a great help, and difficulties that seemed puzzling at the beginning solve themselves readily later on. By our calculations we arrived at the following formula : -.} S — AI Di 2 in which D =: Degree Serivalor, M = Finest size of the tested bale, S = Average size of the tested bale. The following table may serve for practical pur- poses : Not to he used a S° of fine threads inferior to : 1.5 for single warps !)/ll 2.5 " " " 10/12 3.5 " " " n/13 4.5 " •' " 12/14 5.0 " " " l;V15 and thicker. 74 S E R I V A L O R As regards the methods by which till now fine threads were found, there exists none, if the counting of breaks at the winding should not be considered as an endeavor in this direction. Its uselessness will be demonstrated in the next chapter. Chapter V THE ^^M N D I N G 'T^iiF. testing method which we explained in the last -*- chapter is an indirect and tliercfore not an ideal one. What we found out was the size of the weakest spots, what we wanted to know was how resistent or how little resistent this weakest spot will be. From the size we concluded to the resistance, and in a rather reliable way, but the direct way is better than the in- direct one, and to test resistance itself would be no doubt preferable. There exist instruments for this purpose, called dynamometers, of which we shall speak later in the chapter on "Tensile Strength." Not only are these instruments and the whole method of using them unsuitable, as will be proved, but they can furnish only the average tensile strength, which is quite useless, as we may see from this example: A chain A consists of 3 links that, tested indi- vidually, showed the tensile strengths of 102, 83 and 20 lb. ; a chain B, equally of 3 links, showed the strengths of 32, 30 and 28 lb. Although the average :o SERIVALOR strength of A is 68 lb., that of B only 30 lb., A will break at a tension of 20 lb., while B w'ill support up to 27 lbs., that is to say, the strength of B is in reality by 40 per cent, greater than that of A, though the average strength of the latter is more than twice that of the former. Applying, then, the law of mechanics: "A chain has the strength of its weakest links" to silk, we can say : "A zvarp has the strength of its ^veakest spots," and therefore it is this latter and not its average strength that we have to find out. There exist instru- ments also for this purpose, so called "continual dyna- mometers," but they have not proved reliable as yet, and moreover they are working very slowly, so that the testing of 80 kilometers, necessary for this pur- pose, would require days of 10 working hours each. But it appears that well jier formed winding could furnish an excellent way of determining the weak spots of great lengths and an indistinct comprehension of this fact might have contributed towards giving undue importance to the winding, h'or in order to perform it exactly, it would be an indispensable condition that the tension of the thread should be the same through- out — a condition that neither the olficial institutions nor the Laboratory Serivalor are able to accom])lish fully. The difficulties are : 1. The revolving speed of the supplying reel ought to he unvariable during the ivhole operation. This S E R I V A L O R 77 is impossible, as can readily be seen on an inserted dynamometer, which will show an average tension of o(> grams, f. i.. by variations from 10 to 50 grams. •, as they acknowledge in the publication referred to in the first chapter. 4. 7 Jie te)ision during the zeinding must be pro- portional to the sice of the thread tested. This is not accomplished by the Conditioning Houses. They test all sizes by the same tension, about 1 grams, which is absolutely insufficient. A common single coeoon-thread will resist in different j)laces tensions of 8 to IS grams, and the silk thread consists of 4 to 78 SERIVALOR 8 and more cocoon threads ! A tension of -i grams will, therefore, make appear only those threads that lay broken in the skein, and those stuck together threads that were broken by the revolving force of the reel, although they may not even have been weak spots. On the contrary, the Laboratory Serivalor is ac- complishing this task by testing each size under a ten- sion proportional to its average tensile strength. In this way it finds out (with the restrictions deriving from par, 1, etc.) the number of spots that, in a length of 80 kilometers, are thinner than 25 per cent, of the average size, the results being "constant." This is not the place for entering into the details of construction of the apparatus by which the tension can be regulated according to the size. We confine ourselves to saying that by changing the speed and by cautiously employing "rope-friction'' every varia- tion may be obtained. After having got so far, the greatest difficulties were overcome and it was easy to establish the extreme values of : B for the best silk (S' 1) and W for the worst silk (S° 10) expressed in breaks within 80 kilometers. D being the Degree Serivalor, C the number of breaks in 80 kilometers, N the number of degrees desired, we have the formula — S E R I V A L O R 79 c W-B D = — X-1 This classification, together with that of Fine Threads (Chapter IV.) furnishes a rehable indication about the quantity of really weak spots— that is to say, possessing a tensile strength of less than 25 per cent, of that of the average size — contained in the bale tested. The classification of "Winding" must not be con- sidered, how^ever, as more than it is : an indication of the speed allowed for rational winding with 36-10 reels to each girl, keeping reels and girl continually em- ployed. The S° 1 indicates that the bale may be wound with the speed of meters in the minute, equal to twelve times its size; therefore size 13/15 with 12 x 14 = 168 meters. The other degrees indicate: S" ■-. ''^' 4, 5. 6, 7, 8, 9, 10. Size multiplied ])y 11 10 9 8 7 6 5 4 3 that is, for 13/15 meters: 15-4 140 ]:.>(i 112 98 84 70 56 43 But the zvinding does not allozc any conclusions to the other qualities of the thread. That this erro- neous conclusion is pretty general, is in consequence of the fact that not very long ago the weaver was employing only thrown silks for whose quality he ac- cepted the throwster's judgment; and for the throw- ster, of course, the winding forms a very important 80 S E R I V A L O R part of his work. Another reason is that winding is the operation whose results are hrst known to the manufacturer, and this caused the false belief that these results are characteristic of the whole bale. But size 18/20, f. i., is always winding well — does this signify that all bales of this size are of the same quality ? Japans in general are better winders than many Italians. Are they better for that? China Double Extra winding worse than Italians — are they really inferior? Besides, is the judgment of the winding- girls constant? Anyone who will make experiments in this regard will find that they are contradictory to each other, and that the same girl will judge differently of the same bale to-day and to-morrow (of course being ignorant of the fact that it was the same bale). The exact test for winding is important for the throzvster ; for the manufacturer it is but an interesting detail, icithout great value for recognising the general quality of the bale tested. It is true that it is the duty of a good reeler to furnish silk that winds well, as already expressed by the tenth rule of Chapter II. The means to arrive at this end are not generally known, and it is not our object to explain them here. The Japanese have adopted the most radical one : Rerecling. which, however, is too expensive in Europe. Also the "rubbing off" of the hard passages is practiced with good success by many reelers. Some also try to hinder the sticking to- S E R I V A L O R 81 gether of the threads hy applying diluted greasy suh- stances to them before they get on the reel. But this induces to "charging" the silk, and moreover brings about a certain mouldy odor, if not very cleverly done. But it is a fact that there exist good winding silks which are neither rereeled nor greased, nay, that these are the natural product, if the reeler knows his busi- ness. This is sufficient, and those who are producing badly winding silks must bear the consequences. One of the consequences of the sticking together of threads is the Double Ends which occur so often, and which the reeler does not know how to avoid. In fact, they are rarely the latter's fault, as everybody will acknowledge who is acquainted with the work of reeling. The double ends are formed them- selves during the winding by the stronger thread tearing away the weaker one at the spot where they stick to- gether, and drag it along for a length of time, until another break interrupts the double thread ftn-med in this way. (Such passages turned up as extremely heavy ones in the 200 skeins that we tested for Regularity.) Of the same origin are the "underslipped ends" (ends covered by the running thread on the l)obl)in ) which so often trouble the winder. Micropliotograph of a "Flock." A Chapter VI FLOCKS FTER having thus far examined those defects that are measurable by their length, we arrive at those that are measurable no more, but become visible and therefore numerable. They appear as knobs and knots in the thread, which however do not consist, as many believe, of an adherent alien material (waste) but of a normal cocoon-thread that fell off the cocoon in many loops at once and had no time for stretch- ing itself before it was united to the main thread. The opposite photograph (taken from a publication of the Laboratory of the Stagionatura Anonima, Milano, with kind permission of the editors) gives a good pic- ture of one of these "flocks." The origin of this defect was explained in the fifth, sixth, seventh and eighth of the reeling-rules, and we might say at once that it cannot be completely 84 S E R I V A L O R avoided. Good reelers, of course, produce a cleauer thread than careless ones, but it is impossible to pro- duce a thread as clean as it is required by the loom. The flocks must be removed, therefore, from the skeins by special "cleaning-girls," and this is a wearisome task that does not give satisfactory results and, more- over, very often is the cause of the skeins "falling in layers.' The real cleaning must be done by well performed warping. This should be done : 1 . With not more than 300 threads at once ; 2. By one clever working girl and an assistant ; 3. The mounting should be such that the work- ing girls can overlook a length of two yards, and that between the single threads there remains a distance of about J4 inch. 4. The speed must be regulated in order that the working girls may have time for removing the flock and reknotting the thread, without stopping the machine. (Such stoppings produce stripes of dift'erent tension in the war]), which in the tissue ajipear of diflterent luster.) Only such silk ought to be declared as not clean whose number of flocks is too great for allowing their removal without stopping the machine. All others will yield, by this procedure, a cleaner warp than can be jirodnred by the best reeler, and the costs of S E R I V A L O R 85 the cleaning are slight, even in America. By adopt- ing this system, the manufacturer has the possibility of being more indulgent with regard to cleanliness and consequently buying at a cheaper price. The L. S. calls such defects ""flocks" when they increase the diameter of the thread, distin«-uishine however between small ones, that is, those that pass a reed of 58 splits, of No. 9-J:, to the centimeter — 157 to the Paris inch, and bigger ones — viz., those of a diameter below or above 1/10 millimeter. They are found by passing the "-^0 skeins over a dark background and their number is brought into proportion to the kilogram. The extremes for B (best) and W (worst) have been established empirically, but it appeared unfeasible to establish the gradation by arithmetic division of the series ; it was necessary to recur to geometrical progression, according to the following formula : a, b, c, d, e, . . . . being the componenls of the series, and y + z the difterence between a and b. then c = I) + y -r 2z (1 = c + y -f- 3z e = d 4- y -f 4z, etc. There does not exist an official system of judging cleanliness as yet. The silk inspecteurs, however, give their attention chiefly to this quality, accepting "clean" bales as good ones, and "unclean" bales as bad ones. By this they are committing the logical error 8(i S E R I V A L O R of concluding from the parts to the whole, instead of the opposite way. It is the same as if from the fact that every bird has two legs I would conclude that a man standing before me must be a bird, because he has two legs. The inspecteurs, however, reason in this way, thinking good silk is clean, consequently clean >ilk nnist be good. Cleanliness is one of the qualities of good silk; it may be considered of more or less importance by one or the other, but in no case is it considered as Quality itself. We have also seen that it can be improved by subsequent procedures while as Quality can be considered only those intrinsic characteristics that are unalterable. As practical hints for the manufacturer we might add that according to our exi)erience the S° 3 of cleanli- ness is sufficient nearly for all reasonable claims ; but it is not too difficult to find S° 2 and even 1><2. On the other hand even S° 5 might be improved, in the wav explained before, to S" 1. with slight costs. ClIAPTEK VII L O O P S OUR way through the different forms of the original defects of silk, which, as our readers will have noticed, leads from the greater to the smaller, has now arrived at those that are hardly visible and therefore diffictdt to lind. Nevertheless it is necessary to find and control them, for just as the bacilli to mankind, these little defects become dangerous to the loom by their enormous number. The defect which is the object of the present chapter is but a smaller form of the "flocks" treated in the preceding one, but it occurs on the average 5.000 times as often, and therefore makes itself very much felt. \\'hen the roundabout way made l)y the "'bava." before uniting itself with its neighl)ors. is not great enough to render the thread setisibly thicker, we call it "loop." The following schematic design shows un- der a. b, c. d, c. /, (/. the transitions from flocks to 88 SERI VALOR loops ; the gradations are countless, but the line of demarcation is given, as we said before, by the sensible thickening of the diameter. As the loops do not sensibly increase the diameter of the thread, they pass unnoticed also through the finest reed, and in this regard the weaver need not pay any attention to them. But while being closely pressed together on the wari)-beam they get entangled with their neighboring threads and hinder the clear opening of the shed, which is the cause of many breaks and weaving-defects; the weaver therefore justly fears the dangerous little enemy. Its origin is heedlessness with regard to the Seventh of the reeling rules, the complete observation of which, however, is not pos- sible. The worm crosses the thread in infinite varia- tions of density, and it might be said that in this regard not one cocoon is quite the same as another; conse- quently the assorting according to the texture is possible only to a certain degree, that is to say, the defect of loops is inevitable. In fact even the best silk contains about 50,000 of them to the kilo, the SERI VALOR 89 worst, however, more than a hundred times as many, that is about five milhons. The cocoons of widest texture are called in Italy "Bonibaggiati," an inappropriate term, because of not expressing the real thing ; more suitable is the French "Satines" as the wider texture causes greater luster on the cocoon as well as on the tissue. This, our theory of the origin of looj)s, is un- known to the reelers and is published here for the first time. We must not forget that the whole procedure in spinning has not got very far yet beyond the simple principles of a rustic home industry and avails itself very little of scientific methods. That the "Bombaggiati" must be eliminated is known nearly to all reelers — which does not im])ly that they are all doing it — but when I tried to find out what "Bombaggiati"' really are, nobody could give me a precise definition, as they are not judged by clearly \isil)le marks, but found out by an uncertain, instinc- tive distinction. TIow reliable this way is mav be as- certained by the following experiments : Give an order to a very clever "sorting-woman" to clinu'nate all "Bombaggiati" out of a basket of unassorted cocoons ; after a while she will bring back the basket of '"puri- fied" cocoons, and we put it aside. The next dav we give the same cocoons to the same woman in another basket, and she will find more "Bombaggiati" among them, and so we might repeat the thing five times and always nev.' "Bombaggiati" will turn up. 90 S E R I V A L O R Or we take ten cocoons of middle texture (they are quickly and surely discernible under the micro- scope) and show them to ten experts with the question whether they are "Bombaggiati" or not; half of them will answer in the affirmative, the other half in the negative, and none will recognize the real thing : that the cocoons hold the middle between the good and the bad ones in this respect. Even real "Bombaggiati," however, furnish a good thread, if they are not mixed together with others, and are treated according to their nature ; for the procedure adapted to them is not fit for cocoons of denser texture. An official method of testing exists as little for "loops" as for '"flocks." Inspectors think they can judge of the defect by straightening the skein so that it 'forms an even, glossy surface, while looking at it in a corner with their backs to the window, so that their eyes receive the light reflected by the silk. In this way they can examine the surface only, which of course is insufficient, as the inside of the skein remains unknown and. moreover, they are deceived by the circumstance that the loops become the more visible the more lus- trous the silk is. And as it is the task of the good reeler to produce a thread as lustrous as possible, they will see more loops in good silk than in inferior silk. I myself produced once, by a new system, a skein of extraordinary luster, but all the reeling-girls, the fore- woman and the director of the establishment declared SERl VALOR Dl it to be very "downy," that is to say, full of "loops.'" Tested in my laboratory it proved to contain only 20,000 loops to the kilo, that is. better than S° 1. The loops must not be looked for in the skein but on the single thread, where they can be counted, which, however, is not an easy task. The generally used "black" boards, reels, etc., are not black in an optical sense, but dark blue, or green, or brown and very fatiguing to the eye. Only exact optical contrivances allow continual and easy working and clear discerning and counting. Especially do the man}- "casts," viz., beginnings of new cocoon-threads, lead to errors, as they are gener- ally accompanied by looi)s. which however must not be considered as a defect. Dark days and artificial light are unsuitable for the work, and in the Winter, when there are only a few hours of good daylight, these must be well used. The calculation of the ten degrees Serivalor is done according to the formula given in the last chapter, but it is to be observed that the visibility of looj^s is proportional to the square root of the thicker size, by reasons known to geometry (diameter of cylinders). As to practical use, our experiences have proved that the manufacturer may employ even S° 4, but it is safer not to go beyond 3)/j. Of course, the claims are dififerent according to the articles manufactured and to the customers. It is interesting in this regard to con- sider the relation of Cevennes silks to their customers. 92 S E R I V A L O R These silks are known as very "duveteuses," but people simply say, shrugging their shoulders, "C'est la nature de ces soies," which, however, is only partly true. In the nature of these silks lies only a small part of the cause, viz., their thick and hard gum, which, as said in the Fifth spinning-rule, opjioses resistance to the straightening of the thread. The chief cause lies in the fact that French reelers with their system "() la Chauibon' arrive at too scarce a i)roduction and in order to make up as well as possible for this drawback are obliged to a forced speed (about ISO meters in the minute, instead of llO-loO meters with the system "a la tavclle") which does not allow the necessary time for the right soaking and straightening of the thread (see Fourth, Fifth and Fighth spinning-rules). But with the system "a la lavclle" of Cevennes-cocoons there could be reeled a thread that would get S° 1-2 for "loops." Nevertheless Cevennes silks, as they are, have their faithful friends, who even pay good prices for them. This seems to contradict the assertion that "loops" are a serious defect ; in reality it only proves the truth of what we said in the Prospectus : that every silk is good, if emplo\ed in the right way. Silks reeled "() /hall see that it is the case with both. Still more unsuitable appears the establishing of the breaking lengths, if we compare those of textile threads with those of similar forms in other materials. SERIVALOR 113 Steel wire. f. i., has a tensile strength of 120 kilo- grams to the transverse section of one millimeter square, and as its specific gravity is about eight, its breaking length is ] 2,000 =15 8 X 1.000 that is to say, not even half that of the silk thread! Nevertheless, everybody will justly consider a wire rope as infinitely safer for a cable-railway than a silk rope. Considering, however, the three materials: silk, cotton, steel, with regard to their resistance against friction, we learn that steel surpasses the other two many thousand times, while cotton is still consider- ably more resistant than silk. By this we see that the testing of friction is by far a truer exi)ression of the molecular consistency than the deceptive breaking length. On the other hand, the testing of steel wires with regard to their tensile strength gives valuable results ; why are they wrong onlv in our case? Oil account of the ductility of the silk thread. In testing tensible materials, the time of their exposure to tension is decisive, while the instruments actually in use, the dynamometers are completely re- gardless of time. A steel wire that will break under a weight of 114 SERI VALOR 120 kilograms will resist for weeks a weight of 110 kilograms, because it is nearly not tcnsible; but a silk thread breaking under sixty grams will break also under thirty grams if to these thirty grams is allowed the time necessary for extending the thread. The sixty grams are wrongly indicated by the dynamo- meter, because it made the thirty grams increase too rapidly, and before we had time to recognize that they would have sufficed to break the thread. But while the dynamometer makes a rapidly in- creasing weight act only for seconds^ the loom will act the opposite way : it exercises a slight tension dur- ing a long time (several weeks) and every bit of the thread receives on its way from the warp-beam to the cloth-beam about 6,000 jerks, which, though each of them lasts less than one-half second, sum up to a considerable time : about an hour. Seeing, then, how fundamental the difference is between the practical work and the testing instru- ment, we need not wonder that the indications of the latter are of no avail, even if the instrument itself be exact, which a pendulum-dynamometer cannot be, as the pendulum very often receives a swinging start at the decisive moment, which makes wrong results appear. It is possible to correct these to a certain degree, but it requires a skillful and experienced hand to do so. The same with other difficulties, f. i.. that any difference in humidity influences the tensile strength. S E R I V A L O R 115 while it is impossible to bring the thread to a stand- ard humidity. This must be overcome by calculating corrections for each occurring degree of humidit)' (6 to 14 per cent.), by no means easy work. Having finally succeeded in establishing, out of a long series of correct breaking lengths, the values of B (best) and W (worst), we are puzzled by the fact that the best results are given by white silks, while we know from experience that yellow silks are of superior (juality. as is also confirmed by the testing for cohesion. Also this contradiction can be explained, but it would lead too far to follow the long way neces- sary for this jnu'pose. The final result of our researches is: Excellent breaking length is only a proof that the thread was well stretched in reeling (fourth rule), whicli, how- ever, is not sufficient for producing a reall_\- excellent thread ; the essential condition for the latter is good cohesion, which can be arrived at only by observation of the ninth spinning rule. If, therefore, a silk thread shows: a. Good breaking length and bad cohesion, the first is to be considered as deceptive. b. Breaking length and cohesion of the same de- gree ; then the former is superfluous. c. Bad breaking length and good cohesion ; only in this case the breaking length would become inter- esting, for reasons to be explained in the next chai)ter. 116 S E R I V A L O R But it seems that this case does not occur ; at least as far as our experiences go. If further experiments should prove that it never appears in reality (and it is hardly conceivable by what class of cocoons or what method of spinning it should be brought about) we shall abandon the testing of tensile strength as we have abandoned that of ductilitv. Chapti.k XI THE R E S U L r A X T WE IIAV1-: now tested all the various forms of the original defect of silk, the bale is completely analyzed, and having examined the seven indications of Degrees Serivalor, the buyer cannot be in doubt any more for which ])ur])ose the l)ale may or ma\' not be employed. Thus the technical side of the (luestion is solved, but not its commercial side, l-'or the seven hgures of S' are often in contradiction to each other, and in order to express the commercial value of tlie bale, to .say whether it is "extra or "first order," etc.. it is neces.sary to reduce those seven ligures to one. which we call the "resultant." At the first glance this seems verv easil\- done by taking the average of the seven comjionents. But supi)osing an extreme case. f. i.. that the components were: 1, 1, 1, ], ]•, ], 10, they would give the average of 2.20, that is to say, a bale that is extremelv l)a(l in 118 SERIVALOR one regard (f. i., flocks S° 10) would turn out as Resultant 2.3 nevertheless, viz., as "Extra." It is evi- dent that this would not do. Proceeding on this line, we soon find out that the loom is sensible only to the bad qualities, accepting the good ones as granted, just as a man whose one tooth aches does not heed the fact that his other teeth do not ache. Xow we are tempted to jump to the opposite extreme and to say: A bale which shows S" 10 only in one regard, must be designated by Resultant 10. But then we must ask ourselves : Is such a bale really of the worst quality possible, as would be expressed by Resultant 10? What would then remain lor the fol- lowing cases : 1. 1, 1, 1, 1, 10, 10, or 1, 1, 1, 1, 10, 10. 10, and so on, up to 10, 10, 10, 10, 10, 10, 10, which last exami)le only would represent the worst pos- sible quality, from which the others must be dis- tinguished by their Resultant? Reflecting on the matter we arrive at the con- clusion that the influence of each of the single degrees on the Resultant must be pro])ortionately the greater, the worse this degree is. The mathematical way in such cases is to divide the sum of "powers" by the arithmetical sum of the degrees. There remains to find out which i)ower should be employed : various experi- S E R I V A L O R 119 ments have proved that only the second power can be right, and consequently we employ this one. Considering that there can hardly exist a bale of silk that would get S° 1 on each of the seven points of testing, we see that Resultant 1.0 will never occur. In fact, degree 1..") is practically the best and very rarely occurring Resultant. On the other hand, degree 10.0 is nearly impossible as well, and Resultant 9.5 is practically the worst that exists. Between l..~) and 9.5 lay eighty tenths of degrees, which with regard to their "constancy" (see next chap- ter) are reduced to forty, and by this we are arrived at the fort\- qualities of which we s])oke in the Pros- pectus. How to calculate the commercial value of these forty qualities will be ex])lained in another chap- ter. Here we follow with a table showing the relation of the usual designations of "Extra," "Classical." "first order," etc., to the degrees of the Resultant. Be- tween "]^,xtra" and "Classical" there is a considerable difference in price. In the following table resultant 2.5 is still -Extra." while 2.6 is "Classical." The real dif- ference in value, however, between 2.6 and '2.o is not greater than that between 2.5 and 2.4, or 2Ai and 2.T. etc. The degress of Resultant indicate, therefore, as will be explained in one of the next chapters, the real commercial value which is not ex])resscd by the names of the "chops" even if they are genuine, which is not alwavs the case. 120 S E R I V A L O R In order to deserve their designations, the ''chops must not receive a worse degree of Resultant than indicated in this table : Resultant up to So 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 S.5 9.0 9.5 Europe: Extra Classical 12 3 Cliiiia: K.\tra 1 2 3 Double Japan: E.xtra Extra 1 1 ll!/i 1^ 1^-2 2 2^ Canton: Extra 12 3 Chapter XII THE CONSTANCY OF THE DEGREES SERIVALOR As KAcii of the seven Serivalor testings is based on other facts and other methods, the "Constancy" of their results cannot be the same. The differences, however, are not great enough to make it neces- sary to estabh"sh a special table for each of them. It will be sufficient to state the average constancy as follows : If a hundred testings of the same bale would give the average of S' ;5.() in one of the seven items, then; about 20 of these tcstinj^s will yive S" :;.() .iO •' 0.1 above or l)clo\v 1.-) " •■ 0.:{ .. 10 " ■• 0.4 " The Re sultant is still more constant. Under the same suj)posilion, of a hundred testings: 122 S E R I V A L O R 25 will give 5° 3.0 50 " " " 0.1 above or below 25 " " " 0.2 " The average uncertainty of the resultant is therefore 25 X — 50 X 0.1 — 25 X 0.2 = so 0.1 100 which deviation occurs as often toward one side as toward the other, so that it appears neutraHzed. It is better, nevertheless, to reckon with at least half of this uncertainty and to keep in mind, there- fore, that Resultant 3.0 might be considered as being also 2.95 or 3.05. By this the gradation of eighty tenths of degrees, set forth in the last chapter, is re- duced to forty, corresponding each to a difference in value of about five cents, as said in the Prospectus. The question of value will be treated more ex- plicitly in the following chapters. Chapter XIII THE DIFFERENCE BETWEEN THE MERCANTILE AND THE REAL VALUE 'T^ifESE two generally are considered as identical, but -'- . they are so only for the reeler and the dealer and not for the manufacturer. To those who are producing or buying silk, the "real value" is given by the possibility of sell- ing with proht, and as this possibility is dependent on the cost-price, the "real value" is fotuidcd on the latter. It is different with the manufacturer, who can sell only after having produced a new article of the raw material ; for him the possibility of selling with profit is dependent on the fact whether the new article is well made or not. Here, then, the "real value" is connected with the way in which the material is adai)ted for its purpose and with the final result it l)roduces ; the "mercantile value," that is, the cost-price, is of secondary importance. 124 S E R I V A L O R For example : In order to produce a good and comparatively cheap plush I must employ for the pile a glossy and well-covering material. The best in this regard are "Bengal" and "Canton,'' and consequently they have the greatest "real value" for me, much greater in this case than "Italian Extra," f. i., although the "mercantile value" of the latter is much higher. Only after having re- solved to buy one of the said sorts, I begin to take interest in their "mercantile value." For though in this case their "real value" exceeds that of Italian "Extra" I would not pay for them the same price, of course, knowing that 1 can Ijuy them much cheaper. On the other hand, having to produce a Grege-Otto- man with forty-five splits, four threads to the centi- meter, the "real value" of "Canton" is naught, as it is absolutely unfit for this purpose ; but its "mercantile value" remains unaltered because of this. The mercantile value of silver is subject to great changes, while that of gold remains rather constant ; ])ut their "real value" would be naught on a desert island, where some dates could preserve my life and therefore would represent the greatest "real value" for me. Hence follows : I. The "mercantile value" of silk results from the comparison of its (|uality to its price; its "real value" from the com|)arison of its quality to its em- ployment. S E R I V A L O R 125 2. Those who are ignorant of this employment (reelers, throwsters, dealers) are unable to know the "real value." 3. A testing system that establishes Quality al- lows the fixing of the "mercantile value," and, in- directly, also of the "real value," which, however, differs according to how the material is employed. ClIAPTI.R XI\' THE MERCANTILE VALUE THE mercantile value of an article whose price changes daily can, of course, he indicated only relatively, that is to say, in the following way : A hale whose Resultant he, f. i., 4.5, is w^orth X per cent, more than the day's quotation for Resultant 5.5 (Japan 1 1/"-^). This X in reality is equal to 5 per cent., that is to say, each degree of Serivalor is equiva- lent to 5 per cent. With the aid of the following tahle the worth of each Resultant can be calculated, after having ascertained hy how many per cent, the quota- tion of f. i. Japan 1 1/2 differs from the value indi- cated by the table for its Resultant 5.5 : $3.TT. These prices represent at the same time the aver- age ciuotations of the last fifteen years. We see that the value of S'^ 4.1 is doll. 1.0 1, that of S" 4.3 doll. S E R I V A L O R 137 4.00 ; the intermediate two-tenths of a degree are to be considered only as one with regard to "Constancy" (see Chapter XII j the value of which is four cents, as stipulated in the Prospectus. RELATIVE MERCANTILE VALUE OF THE RESULTANTS, IN DOLLARS. (Tenths of Degrees.) )eg. .0 .1 .2 .3 .4 .5 .6 . 1 .8 .9 1.. 4. 58 4.56 4.54 4.52 4.50 2.., ..4.47 4.45 4.42 4.40 4.38 4.36 4.34 4.32 4.30 4.28 3.. ..4.26 4.23 4.21 4.19 4.17 4.15 4.13 4.11 4.09 4.07 4.., ..4.05 4.04 4.02 4.00 3.98 3.96 3.94 3.92 3.90 3.89 5.., ..3.87 3.S5 3.83 3.81 3.79 3.77 3.75 3.73 3.71 3.69 6.. ..3.67 3.05 3.64 3.62 3.60 3.59 3.57 3.55 3.54 3.52 7.. ..3.50 3.49 3.47 3.45 3.44 3.42 3.40 3.3S 3.36 3.35 8.. ..3.33 3.32 3.30 3.29 3.27 3.25 3.24 3.22 3.21 3.19 9.. ..3.18 3.16 3.15 3.13 3.12 3.10 The great difference in value between, f. i., Resultant 4.0 and 1.5 might call forth the question whether this difference is justified by the conditions of production. The answer to this question is not quite simple, it is, directly, no, indirectly, yes. From the direct point of view, that is to say. con- sidering the higher costs resulting from the better quality of cocoons and the greater care of spinning, the price of "Extra"' appears exaggerated. The reeler, however, does not know his own cost-price. It is bv no means easv to calculate, neither is the reeler 128 SERIVALOR very much interested in it, seeing that his selhng price is not dependent on his cost-price but on the quota- tions of the market. This makes nearly all reelers conceive the strangest ideas about their cost-price, and there are not two among 100 of them who would agree on this point. After long studies the author has come to the conclusion that from the changing wages, price of coal, and worth of the residual products, there might be calculated an average of lire Jj.OO to the kilo (twenty-five cents to the pound) for wages and general costs, if the selling expenses are not too high, that is to say, if the reeler does not, f. i., keep a special office in Milan for selling a small i)roduction, or if he has not to pay too high interests to the Commissioner for the advancing of money for buying cocoons. In order to tind out the ditlerence between the cost and the selling-price, it is necessary to compare the averages of the quotations of cocoons on the one hand, and of silk on the other, during a long period. This difference gives the selling price of the reeler's work, viz., what costs him lire ."l.OO. This selling price was, on the average of the last 10 to 15 years, lire 3.50, that is to say, the average gain of the reeler, who is exposed to a thousand risks and perils, was up to 1914-15, 50 centesimi to the kilo, that is four cents to the pound. If we consider that the Italian reelers, whose number is about a thousand, had paid the crop of 1914 about K5 millions of lire, and that, by the S E R I \^\ L O R 129 collapse of prices caused by the great war, they have lost at least thirty millions, we must acknowledge that in general they are not in an enviable condition. And now we are arrived at the "indirect" side of the question, whether the high price of "Extra" is justihed. The great majority of Italian reelers get on with but small capital, slight credit, at times high interest, and have no constant customers. They do spinning for stock, and as thc\' cannot obtain good prices, as nobody would give them credit for producing "Extra," even if they did so, they are always working at great speed, trying to arrive at a high production under any condition, that is to say, they are producing qualities of Resultant 4.0 to 5.5, while many among them would be able to produce Resultant 2.0 to 3.5, if buyers would only allow them one to two lire more to the kilo (eight to fifteen cents per pound). Nay, they are not only as able to sj)in as ex- cellent silk as the most famous "Marca" reelers, but they are in better condition to dtj so, as they quite gen- erally manage their estal)lishments themselves, to- gether with wife and children, while the "Marca" reeler generally lives in Milan, rarely visiting his establishments and completely relying on his staff for their managemeiu. Thus we see, on the one hand, a majority of hard working people who are scantily recompensed for their labor, on the other, a small minoritv of wealilu" men 130 SERIVALOR who are independent enough of the market that they can refuse to sell if the prices of the day allow them too small a prolit. They consequently demand not only the advance of one to two lire justified by their higher costs (some of them believe this difference to be twice as great), but they claim also the legitimate profit of their tiresome and risky work, of which the small reeler is robbed by his helpless position. In this indirect way we come to the conclusion that the higher price of the better (jualities is not unjustified. In spite of this it is ])ossible for the dealer or the manufacturer to buy good silks at a price only from one to two lire higher than that of "1st order" if, recognizing the rceler's ])()siti()n, he will treat him cleverly and benevolently, and also allow him the few weeks necessary for improving his products with the help of a good testing establishment, and if, after having arrived at this point, he keeps him constantly occupied w'ith his orders, so that he may not be com- pelled to work on stock again. It is necessary, moreover, not to change too often the sizes ordered, and not to demand a size which the reeler cannot produce, because the cocoons reared in his neighborhood do not yield it. The richer reeler is able to defend his interests in this regard and does not accept a size, for which his cocoons are not fit ; the poor one is often compelled to give in. His cocoons have, f. i., a "bava" of 2.8, and he knows SERIVALOR 131 that five of them yield a good l'^/\i^. Now the buyer orders i;5/l-^ : instead of refusing the order he arranges a way of spinning and gives the order to com- bine three big and two small cocoons, that is to sa}-. he deliberately acts against the second spinning rule, and the result is bad silk. Another expedient lor him would be the well-known twofold sizing, but also this is less practicable I'or him, because he cannot run the risk of having the l^ales refused if the trick should once fail. Finally the buyer must not induce the reeler to speculations, demanding of him to accept orders on long delivery in a moment when prices arc low. The rich reeler is leading a continual war in this respect with his buyers, a war which in the course of \ears brings about as many victories as defeats for both parties, but wliich embitters them both. In the end both of them have hit it as often as not, and the out- come of the long silent fight is that neither has won anything. P)Ut in the rich reeler the buyer at least has a solvent o])ponent who can i)ay when he loses ; but what can he claim from a ]:)oor man? If he has succeeded in enticing him to accei)t a disadvantageous contract, he only compels him to buy bad cocoons, in order to save himself from ruin, if possible, and consequently to produce bad silk. The latter will be recognized as such by a reliable testing establishment, but it has not become better for this. The buyer has the right to cover himself on the reeler's expenses. 132 S E R I V A L O R but it is a hard thing to obtain payment of the dif- ference from a man of scanty means. The right way of concluding contracts is to base them on the official quotations, for instance : The reeler nuist deliver a (|uality not worse than Re- sultant '3.0 ; the price to be paid is calculated by the average quotations for "1st order" of the last three months preceding the day of delivery, with an ad- vance of 1 to 2 lire. An\l)()d\- who will com])are the prices of his purchases of the last five years, f. i., with those he would have obtained by this method, will find that on the average he would have paid less. He would have saved himself the trouble of hitting the right moment and all the excitement connected with this system, and instead of the usual dift'erence of three to five lire between "Extra" and "1st order," he would have paid only one to two lire. There are very important manufacturers who have been bu}ing according to this system for many years, and they evidently find it profitable, for they stick to it. If the 'advantages of an absolutely reliable test- ing system are great with regard to the Mercantile- Value, they are still more so with regard to the Real Value, which will be the subject of the next chapter. Chaptkk X\' THE REAL VALUE n^iiE author is not infornied in regard to the -*- American wages for winding, and heing driven into exile by the war, it is at i)resent impossible for him to get exact information. He will therefore try an average calculation. Suppose a winding-girl working with forty reels at a speed of 150 yards the minute, and arriving at (K) per cent, effective production, is i)aid sixty-rive cents for ten hours of work; she will produce 1540 grammes of ]:)/15 in ten hours, and consequently the wages would be ecjual to twenty cents per pound. In no case can this supposition diff'er far enough from the actual to seriously aft'ect the following con- clusions. Generally there are : The wages per lb. for warp- ing arc 5l/> times those for winding; the wages per 11). for weaving are 10 times those for winding, and the general expenses, including selling expenses, twice 134 S E R I \' A L O R the wages for weaving. So we have : Winding 20 cents, warping 5U cents, weaving $2, general expenses $4 per pound. All these wages and expenses ought to rise and fall in inverse proportion to the produc- tion, that is to say, if the material yields to a double production the wages ought to be reduced to a half. With daily pay this reduction is brought about automatically, but with piece-work the adaptation is not quite so easy, and therefore we will leave this factor aside for the present. But doubtlessly the re- duction is brought about in the general expenses, and these will be diminished therefore by twenty cents per pound if the production is increased by 10 per cent. Now it is ascertained by experience that an in- crease of the production by 10 per cent, is caused by employing, for the same article, a quality of silk (raw or organzinc) by one degree of Resultant better than the one employed before, as, f. i., Resultant 3.0 in- stead of 4.0. Therefore tJie Real Vahie of one degree of the Resultant is, by diminishing of general expenses alone, about tzi'enty cents. Besides there ought to be also gained, in the course of time, three-quarters of the difference in wages for piece-work, while the last ({uarter ought to go in favor of the workmen. For by the constant employing of good silk also the wages for piece-work may l)e reduced without complaint of the workmen, if the thing is done l)enevolentl\-, that is to say, in a way S E R I V A L O R 135 that at the end of the year the weavers have earned more than before. This can be easily brought about by the said system especially on the occasion of fixing the wages for new articles. It is true that it can be correctly done only if the wages are calculated "a i)riori" quite justly and exactly, a task to which not all managers are equal. Perhaps that later on I shall publish a system of such calculations. In my own experience as a manager the results were : After three years, and after having reduced the 135 hand-looms of the establishment to twenty- four, the general expenses were diminished by $10,000 (^17 per cent.) while the average gain of the hands was increased from $!)5 to $110 yearly (=16 per cent.) and the yearly production of each loom was raised from 3,000 yards to 4,500 yards (= 50 per cent.) with a greater proportion of good (jualities. Each of the 20,000 pieces produced was diminished in cost by $1.50 for wages and general expenses, in com- parison with those of three years before, the wliole profit amounted therefore to $30,000. It is true that this success may be ascribed to the circumstance that the whole organization was rather imperfect, so that there was occasion for many im- provements. lUit the facts set forth here ])rove that two objects ap]:)arently opposed to each other may be obtained at the same time. 7'ic., diminishing of wages for piece-work, and increasing the workmen's era in. i;JG S E R 1 V A L O R It is evident that the workman does not consider how many cents he gets to the yard but only what his pay is at the end of the week. In this regard I had two experiences : When I took up the management every technical improvement was considered as a hostile act ; long years of suffer- ing under narrow-minded management had taught the workmen that every change aimed at a diminution of their earnings, and consequently they had to be com- pelled into every reform or improvement. On the contrary, when I had been there for two years, I might have proposed to the men a diminution of wages by 10 per cent., and they would have accepted it readily, knowing very well that every change meant an increase in their earnings. Regarding the relations between manager and workmen many things may have changed in the mean- time in Europe, and it might have been different in America all along, but the fundamental conditions of the struggle for life are the same everywhere, and therefore from what I said before there results, that by employing the degree of Resultant which with in- creased speed yields the most efficient production, the better degree must turn out to be the cheaper, in spite of the fact that it costs more to the pound. TJiits tJic Real Value of a degree of the Result- ant might be estimated as: 20 cents, and ^ of twenty- seven cents = 30 cents; hut for prudence sake zve zvill say twenty-five cents. S E R I V A L O R 137 Of course the advantage can be obtained only if the speed of the looms is increased proportionally to the better quality of silk and if this higher speed is inainfaijicd constantly ; which again is only possible if every bale has been thoroughly and exactly classilied before. It is not indispensable that this classification should be done by the Serivalor-system, if it is only exact and constant ; but in this case it will needs fol- low the principles set forth in this publication. There is another advantage of a just and im- ])artial testing system, which 1 might call the prophy- lactic one. Just as everybody will prefer remaining well to being cured, it is better to receive a good quality than to be compensated afterwards for the loss on a bad one. The rceler who knozcs that he is under control will in most cases produce good silk. Although he can- not be quite sure of the quality of his product, he, and especially his working people, who very soon be- gin to feel the control, are able to eliminate, by care and attention, nine-tenths of the causes of bad pro- duction. Hut in order to arrive at this, it is necessary that the testing should be performed by an impartial es- tablishment, that is to say, neither by the bu\ers nor by the reelers themselves. The two parties would hardly agree, especially in times of great fluctuations of i)rices, and they would 138 S E R I V A L O R have to refer the matter to an impartial institute, which must be maintained by the whole trade, and which w^ould be obliged to demand high fees if applied to only in cases of dissension. But the moment the public institution exists, every ])rivate testing becomes super- fluous, and therefore all endeavors to invent testing methods which could be emi)loyed by anybody in his own house, without studies and without instru- ments, are useless, besides that they must needs remain unsuccessful, as will be recognized by everybody who has followed me so far. It may be that many of my readers will be under the impression that the clerks of a testing establish- ment according to the Serivalor system must be learned men or at least a M. A. of an university ; nothing of the kind is required, just as it is not neces- sary to be a mathematician for using logarithmic tables. The collecting of the material and the cal- culating of the Serivalor tables required many years of work, and their use demands only intelligence and attention. If private testing were of any use. why is the establishing of the weight left to the Conditioning Houses? Certainly not because the apparatus is too costly or its handling too difficult. A manufacturer who bu}s 100,000 kilos of silk yearly pays, in Europe, 3,000 francs for their conditioning, while the apparatus costs only 500 francs and the work may be done by anv of the clerks. Nevertheless nobodv thinks of hav- SERIVALOR 139 ing the conditioning done in his own house. How much less logical is this for the far more difficult and complicated testing of quality ! Also the conditioning houses have not yet been in existence a long time. I tried in vain to find out the day of their origin in Milan, but according to all information they cannot be even a hundred years old. How was it before? Certainly many people who produced silk at that time cheated as much as they could by moistening. But no doubt there were also honest men who would have preferred fair trade ; these and the dealers who were the injured ones in this case, must have conceived the idea of an official condition- ing institute — an idea the execution of which may have met with no small difficulties. It were certainl)- not the honest men who opposed it — just as to-day there are no honest men who are oi)posed to the idea of an official testing institute — but finally the thing was carried through, and to-day we cannot imagine the silk trade without the conditioning houses. Doubtlessly it were chiefly the dealers, whose in- terests were at stake, who brought about the innova- tion. But no immediate interest either of the producer or of the dealer demands a change in the testing of size and of quality, although they Ix)th would certainly profit from a solid and honest basis for the trade, after the "Testing House" had been in action for a certain uo S E R I V A L O R time. But meanwhile they are not so injured by the present state of things that they should be in a hurry to alter it. They leave the matter to itself and confine themselves to proposing little reforms of usages, etc., to conferences and congresses. In all these assemblies only producers, throwsters and dealers are to be seen or heard, but hardly a manufacturer. Among the more than 100 members of the International Congress of Torino, 1911, there was only one manufacturer; America had not even sent a single representative ; why this ? Because the manufacturer is too much occupied with the cares of production and selling to remember that the surest profit is that made in buying; and he has, as it were, 7io time to defend his interests. And yet the reform of the silk trade with regard to the testing of Quality and Quantity must be initiated by the manufacturer, as it is his interest that is at stake in this regard. In the centers of production : Yoko- hama, Shanghai. Canton, and Milan, there exist only reelers and dealers, but no manufacturers, and it is quite natural that the usages established in these places are in accordance with the interests of those who made them. But there is a town where the manufacturers form the enormous majority and whose consumption is so preponderant that it commands the attention of the whole trade. What is earnestlv claimed bv New York will be S E R I \ A L O R 141 accepted by all markets as readily as there ever was accepted a just and rational reform, and therefore to the author it appears as a duty of the American manu- facturers to take the lead in the question of the Reform of the testing of Quality and Quantity. They may be sure that by doing so they will earn the gratitude of their European brethren, and finally also of the producers and dealers. We have now exhausted our main subject, the right valuation of silks. The following chapters will treat technical questions in connection with silk, but not wilh its value. This reproduction from tine original photograph sliows the dif- ference hetween spht and unsplit ends. Chapter XVI LOUSINESS DYED silk not rarely shows a Hght-hued clown to which was given the appropriate name of "lousi- ness." The causes of this defect remained unknown for a long time. We shall try to give an account of the ;respective researches. The cocoon-thread is compo.sed of two "elemen- tary threads," which on their part consist of countless little fibers. For a long time this structure of the ele- mentary threads was contested, and even L. Blanc in his excellent study (1) has still asserted, with all his authority, its homogeneity. But Italian and German scientific men continued to furnish new ])roofs to the contrary, and finally A. Conte and D. Levrat succeeded in proving the fibrillous structure ( "i ) by especially (1) Etude siir la secretion de la sole. Annalcs du Lab. d'ctudes de la sole. Lyon 1887-88. (2) Snr la structure fibrillaire dc la sole. Annales 1901-02. lU S E R I \^\ L O R conclusive experiments, so that this question appears theoretically solved. Weavers may persuade themselves in the follow- ing way : Decomposing the warp-threads of a piece- dyed stuff under the microscope, we discover, among the elementary threads of various size, some which distinguish themselves hy their very small diameter. Measured micrometrically. their diameters appear as 1/5 to 1/10 of those of the rest, they consequently possess only 1/25 to 1/100 of the hody of ordinary ele- mentary threads, and this fact alone must call forth reflections about their origin. ]\Ioreover. in counting the elementar\- threads, we find that very often they appear odd-numbered, and as the cocoon-thread always consists of two elemen- tary threads, we see that one or several of these must have been split. Following the thread with the micro- scope-needle, we soon arrive at the point where the split-off fiber unites itself with its original thread, and by this we have the proof of its fibrous structure. The photo on i)age 142, taken, by kind permis- sion, from a publication of the Laboratory of the Stagionatura Anonima ]\Iilan, shows the difterence be- tween split and unsplit threads. The splitting, however, occurs in raw silk, and all defects of the latter consist at least of one intact cocoon-thread. It is therefore useless to examine raw silk with regard to the question whether it will become ''lousy" in the dyeing, for this defect is caused solely S E R I V A L O R 145 by split elementary threads. It has nothing to do with the quality of the silk, but is a fault of the dyer's. Although this defect is as old as dyeing itself, it was not before 1806 that it began to be scientifically examined. From the studies published on this sub- ject by the Laboratory of the Milan Stagionatura Anonima under the direction of Professor Gianoli, and by Professor Lenticchia of Como, (1) we gather: (1) Lenticchia: "Sopra tin niiovo difctto delta seta di Bombxx viori." {Boll, di sericoltitra 18 and 25, May, 1S9G.) Gianoli: "Intorno alia iinperfecione degli attuali sistenii di tintura delta seta." (Bol. d. ser. 6, March, 1898.) Gianoli: "Comniunicazione preliminare snlle cause die provocaiio la sfilacciarsi detle scte tinte." (Relatione presen- tata at 4". Congresso di Bacologia e Sericoltura in Torino 1S9S.) Laboratorio delta Stag. an. Milano: "Intorno al difctto di sfilacciarsi di iin filato di seta tinta." {Boll. d. ser. 21 Mai, 1900.) Laboratorio: "Intorno alio sfilacciarsi dcllc scte durante Ic operazioni tiniorie." {Boll. d. ser. 10, March, 1901.) Lenticchia: "Niiove osservazioni ed csperienae snlla fonna- zionc dei fiocchetti nclla seta del filngclto." {Como 1902.) Laboratorio: "Appunti allc osservazioni del Prof. Len- ticchia." {Boll. d. ser. i:5, April. 1902.) Lenticchia: ".Incora snlla formazione dei fiocchetti delta seta." {Como, 1902.) Laboratorio: ".Incora sui fiocchetti dcllc scte." {Bol. d. ser. 30 Nov., 1902.) Lenticchia: "S em pre sui fiocclictti dclla seta." {Como, 1902.) Lenticcliia: "Snlla forma, composizione e strnttura del filo serico." {Milano. 190;!.) Chapter X\'II LUSTER AND COVER ALSO these two qualities can be judged according to the Serivalor system, and our laboratory will do the testing on demand ; but, being without connection, nay, sometimes even in op])osition to the other De- grees Serivalor, they would make appear a wrong Resultant, and therefore cannot be comprised in the latter. For this reason it would also be useless to explain the way of testing them ; it will suffice to giye the following general indications. The luster of the thread and its capacity to cover the tissue are qualities of race and therefore independ- ent of the method of reeling. They are proportional to each other, that is to say, the more lustrous a thread is, the better cover it will yield, and vice versa. But the luster of raw silk has nothing to do with that shown in the tissue. For both qualities the twisting is the decisive factor. Therefore, raw silk can be compared only to raw silk and thrown silks onlv to others of the same S E R I V A L O R 147 twist, but the comparison must never be done in the skein. Dyed siuudtancously and treated the same -z^'ay, all skeins of razv silk zvill shozv the same luster, and so also all skeins of thrown silk of the same twist. But even in the tissue the difference between more or less histrous silk is diminished : (a) The heavier the count, (b) The sharper the torsion is. More lustrous materials yield a cover of high lus- ter in single weaving, and also a more lustrous tram, l)ut with an organzine of ooO turns to the meter, the difference is hardly perceptible any more. The higher luster of the tram will be visible only in light weft tissues, and also with single-warp satins the dift"erence is more striking in forty splits two threads, than in forty splits six threads to the centimeter. Under condition of the highest luster (as with light-count satins), the relation of the extremes is: Fifty threads of S"l have the same lustrous eft'ect as seventy-two threads of S"10, of the same size. But this does not mean that two tissues woven of materials of this proportion would have the same aspect. For this it would be necessary that the cover should be also equally deep. Silk is diaphanous, and the ex]:)erienced eye easily distinguishes the thin, shal- low size from the thicker, deeper one. Therefore, the proportion established above is valid onlv for eft"ects of the surface as dent-streaks, etc. ClIAPTKR XV'III THE TOUCH I.N" GENERAL the touch of the tissue is in inverse pro- portion to its hister. and proj)ortional to the S" of cohesion of the thread. It depends, however, more on the race than on good reeling. China silk yields a perfect, that is to say, rich and at the same time soft, touch ; next to it come : Japans; then Toscana and Turkestan. Satisfactory in regard to richness of the touch, but not to its softness, are: Levante, Gialli puri, Cau- casians, Incroci Chinese. Of firm, but rather hard and rough touch, are : Incroci Giapponesi, Persia. Bengal and Canton yield a flabby but smooth touch. In regard to this quality no gradation of testing is established. ClIAPTI.R XIX THE LOSSES BY PREPARATORY PROCEDURES (A) IV ill ding: The loss by the winding is ]M"oportional to : The carefuhiess of the winding-girl ; the circumference of the reel ; the degree of Serivalor for winding. With a careful girl and a circumference of 1^^ meters, the loss is : For S" Winding : 1 2 ;3 I .■) 7 8 10 Per cent. 0.2 0.1- 0.6 0.8 1.0 1.2 1.1 1.6 1.8 2.0 (B) Cleansing during the icar/^iiig, or throning. The losses are not important : 1.30 SERI VALOR For S" Flocks : ' 1 2 3 4 5 6 7 8 9 10 Per cent. 0.1 0.2 0.3 0.-4 0.5 O.G 0.7 0.8 0.9 1.0 Lenticchia is of the opinion that the defect Hes in the silk thread ; Gianoh that it is caused by the dye- ing. The latter states that flocks occur in every dyed skein, but that they remain unnoticed if their number does not surpass 150 in 1,000 meters of thread. By cautious treatment on bobbins he succeeds in dyeing nearly without flocks a small quantity of silk, of which the rest has come back "lousy" from the dyer. Lenticchia objects that Gianoli's delicate method cannot l)c employed by the industry and asks: "Why, with the same treatment, one lot of silk becomes iousy' and the other not?" He believes the reason to lie in the weak constitution of the thread which may originate from worms tliat have suffered by disinfect- ing vapors. As this method is chiefly used in Italy, his sup]:)osition would diminish the value of Italian silks. Gianoli. on the contrary, proves that Italian silks ,ire not more subject to this defect than .Asiatic ones. Lenticchia acknowledges this fact, but for the rest sticks to his opinion, which he sui)ports by new re- sults of his studies, viz : (1) In the silk thread liable to '"lousiness" the fibrin is not only surrounded, but penetrated by the sericin. S E R I V A L O R 151 (2) Worms treated with disinfecting vapors pro- duce a thread penetrated by the sericin and conse- quently gets "lousy" in the dyeing. (o) The end of the thread towards the chrysalis is flatter and more penetrated by sericin than the rest. Of this I wish to remark: (a) It is true that microscopic flocks appear in nearly all dyed silks, but in order to form a technical defect they nmst be visible to the naked eye. Stretch- ing a dyed skein so that it forms an even, glossy sur- face, and regarding it with his back to the light, one discovers the flocks as little i)ale dots. Their lighter hue is owing to the small diameter of the si)lit-ofl:' fibers which consequently are more diaphanous than the unsplit elementary threads. By the same optic law the foam of a colored liquid ap- pears much lighter than the liquid itself. My explanation was accepted later on by Professor Gianoli in a lecture given in the Chemical Society of Milan. (b) To Professor Lenticchia's question, why, with the same treatment, some silks become "lousy" and others do not, I reply : Just on account of the same treatment. If I should treat a race-horse the same as I might a pack-horse it would perish ; l)ul for this it is not to be considered as a degenerated animal. 152 SERI VALOR (c) I have had dyed dozens of bales of ItaHan trams without finding fiocks ; but I found them very often in white "Levante," and also most of the law- suits concerning "lousiness" were caused by white silks. According to my experience I must therefore consider these silks: (Persia, Turkestan, Brusa) as especially Hable to "lousiness," although there does not exist a sort which would be quite exempt from this defect. (d) That fibrin penetrated by sericin is liable to splitting appears probable, but I don't believe that this penetration is owing only to the causes to which Prof. Lenticchia ascribes it, for the Persians do not use dis- infecting vapors, and if the spinning of the cocoons up to the last bit were responsible for the defect, the latter would occur oftener with Italian silks than it does. ]\Iv assertion that "lousiness" is independent of the quality of raw silk was confirmed by the following experience: I tried to employ Brusa 12/14 for the warp of a satin of forty-five splits, three threads, to the centimeter, in order to see whether it would turn out very streaky. The winding, warping and weaving (width 130 centimeters, KjO strokes to the minute), went on regularly, and the piece came back faultless from the dyer's. The rest of the bale was employed for tram, and this was dyed "lousy" by tzvo important dyers. Not to mention that the testing of the said Brusa had given a good degree of cohesion, its good quality SERIVALOR 153 was made evident by the fact that it could be em- ployed for forty-five splits, three threads to the centi- meter, a count of reed, for which good Canton is unfit, while trams even of the worst Canton may be dyed free of "lousiness." Seeing, moreover, that the same material did very well for piece-dyeing, we come to the conclusion that — (1) Raw silk of good quality may be liable to "lousiness" nevertheless. (2) "Lousiness" is the consequence of a fault of the dyer's. I tried myself to dye the said tram in the labora- tory and succeeded in dyeing it with or without "lousi- ness" at will. As Prof. Lenticchia had said in his article: "I possess a skein of silk; a wreath of laurel to him who is able to dye it free from 'lousiness' in the ordinary way (that is to say. not on bobbins)," I wrote to him to send me the skein, marking it with a sealed string. This string would also prevent the dyeing on bobbins. (In fact, I am employing sticks.) Prof. Lenticchia replied that he did not possess the skein any more, and congratulated me on my invention. I presume, though, that the dyers know very well that the fault is theirs. This supposition seems to be confirmed by the fact that I received only one answer to the Prospectus of the first edition of my book, in which I had announced elucidation on this matter, and 154 S E R I \^ A L O R which was sent to all the dyeing establishments in Europe. As I cannot believe that these establishments do not take interest in new publications concerning their industry, there is no other supposition left but that they did not want to spend money to learn what they knew already. To sum up : There are silks Ti'/z/V//. independent of their quality, are more liable to "lousiness" than others; but even those can be dyed free of this defect. Chapti'.r XX THE DIAMETER OF THE SILK THREAD THE diameter of the same size is different accord- ing to the number of cocoons employed for it. This number varies considerably according to the race : Size 13/15, f. i., may be made of 4, 5, G, T, 8 or even 9 cocoons (Canton). The differences occurring may be calculated in the following way : We suppose a silk thread to be composed of four cocoon-threads (bava) whose diameter be 1.0 and whose weight equally 1.0. If the same size should be made of 5 "have," their weight must be 0.8 (4 X 1-0 = 5 X 0.8) and the diameter will result from the equation : 1.0 : X =|/l.0 : 1/1X8, which gives X = 0.9 In the first case we unite 4 "'have" of diameter 1.0, which give a smallest diameter of 2.0, while in the second case o "have" of the diameter 0.9 are 15(5 S E R I V A L O R united, which must give a diameter of more than 2.0, as can easily be seen by anybody who will group to- gether 4 and 5 disks of these proportions. In fact, the smallest diameter of 5 threads must be 2.34. Putting 100 for the smallest diameter possible, we have for : Cocoons : 4 5 G 7 8 9 Diameter: 100 117 103 113 115 100 Another factor of the diameter is the more or less complete stretching of the bava, of which the quality of the thread is dependent, and therefore it might be said in general : The smaller the diameter of the thread is, the better its quality; but we have seen in the respective chapters how difficult it is to judge by these indications. Apart from these restrictions, there are two ways of finding the average diameter of a silk thread, both by supposing the thread to be a cylindrical body and consequently having a circular transverse section. first method : The diameter of homogeneous cylindrical bodies can be derived from their specific weight. Accepting the silk thread as an homogeneous body, by means of the pykometer, 1 found its specific weight to be 1.2!) to 1.;]0. S E R I V A L O R 157 The diameter (D) of cylindrical bodies being pro- portional to the square root of their weight, D = X K'"— and from the specific weight of 1.3 it results that X = 1.0 (expressed in microns := thousands of milli- meters). (L. Vignon in his "Rechcrchcs sur la densitc de la soic' {Annalcs dii Laboratoirc d'etudcs de la sole, Lyon, 1889-1890) had established a specific weight of 0.9 to 1.1 by the mercury method. A year later, by immersion into benzine, he found l.;53 to 1.31. Vignon rejects the method of the pykometer, as he thinks the sericin must be dissolved by the boiling water. I beg to object to this, that during the few moments that the experiment lasts the solution must be so slight that it influences the correct result much less than the fric- tion of the skein against the benzine, which diminishes the sensibilit}' of the balance. The mere fact that the result obtained by my method is smaller than his will furnish to every physicist a sufficient jiroof that mine must be right.) As, however, the silk thread is not homogeneous, but a bundle of roundish bodies, X must be greater than 10. (The transverse section of a cocoon-thread might be described as two isosceles right-angled triangles with rounded-off' corners joined together. ^Microscopic lo8 SERIVALOR photos are to be seen in : "Notice siir le laboratoire d'etudes dc la soic. Condition des sois, Lyon. Others on a larger scale in Prof. Lerjticchia's "Sulla forma, composicione e struttnra del filo serico," Milano, 1003). Supposing the cocoon-thread to be of cylindrical form, X must be at least as much larger than 10, as the outward square is larger than the inward circle: X : 10 = (2r)- : r- tt Consequently, X must be at least 12.7. But, as said above, the cocoon-threads are not of completely cylindrical but of roundish sha])e, and therefore cannot be joined together as tightly as it would be possible with cylinders. Under the microscope the gaps between them are clearl)- visible. (See Chai)ter V'lII. Cohesion.) From this results that X must be larger than II), and it may be concluded that it cannot be much below 14. Second method: The diameter of cylindrical bodies might be measured directly by joining them together without gaps on a fixed length ; in our case, by winding the thread around a dark board. If a thread of a certain length, in meters, (L) and a certain weight, in tenths of milligrams {W) joined together F times, covers a certain length in microns (S) on the board, then: S y vv L SERIVALOR 159 The winding of the thread around the board re- quires much patience and a certain skill, both of which qualities are not rare with weavers. After some effort we succeed in joining together the threads so that under eightfold linear enlarge- ment they showed neither gaps nor doublings. The flattening of the threads is avoided by their compara- tive hardness. By this experiment we arrive at the following figures : "Piemont" W z= 35 L= 1.125 S = 1560 F = 20 X = 14.0 ••China" W = 77 L = 1.125 S — 2!500 F = 25 X = 13.8 "Brussa" W = 22 L= 1.125 S= 900 F = 15 X = 13.G '•Bengal" \V = 48 L = 1.125 S = 1820 F = 20 X = 14.0 "Bengal" \V = 40 L = 1.125 S = 1720 F = 20 X = 14.4 These being sufficiently in accordance with the results of the first method, we may state that X = 14. (Tn the "Aimalcs du lahorotoirc d' etudes do la soic," Lyon, 18SG and 188T-88, are published many microscopic measurings of the diameter 'of the cocoon- thread. According to these, the value of X oscillates between 14 and 20. But. as it is not said whether these figures concern the larger or the smaller diam- eter, and as moreover the proportion of these two changes the more the thread approaches its end, those measurings do not aft'ect the results we arrived at.) Thus we may establish the following table of diameters : 160 S E R I V A L O R Size 4 5 6 7 8 9 10 Mikron 29.5 33 36.1 39 41.7 44.3 46.7 Ten thousandths of inch = Alikro-inches 11.6 13 14.2 15.3 16.5 17.4 18.3 Size 11 12 13 14 15 16 17 IS Mikron 4S.9 51.1 53.2 55.2 57.1 59 Gl 62.6 Alikro-inches 19.2 2U.1 20.9 21.7 22.5 23.2 23.9 24.6 Size 19 20 21 22 23 24 25 26 Mikron 64.3 66 67.6 69.2 70.7 72.3 73.8 75.2 Mikro-inches ....25.3 25.9 26.6 27.2 27.8 28.4 29.0 29.6 Size 27 28 29 30 31 32 33 34 Mikron 76.7 78.1 79.5 80.8 82.1 83.5 84.8 86.0 Mikro-inches .... 30.1 30.7 31.2 31.7 32.3 32.8 33.3 33.8 The.se figures are important for the weaver, as they indicate the space occupied by the warp-threads in the reed. Another table in the chapter concerning reeds will sliow the space left by the various reeds, and by a comparison between those two we recognize the increasing friction of heavy warps and the neces- sity of the better cohesion, the more threads are pressed together in the diminishing space. T CiiAPrER XXI THE ALTERATION OF SIZE BY THROWING. iiKRi-: are two factors whicli alter the size during the throwing : (A) I'hc l)reaking of the thread (hiring the wind- ing and the following oi)erations : This hreaking generally occurs at the weakest, that is to say. thinnest parts of the thread and is followed by the removal of a certain length ; the thinner parts thus being eliminated, the rest has become heavier in proportion to its length. A rough estimate tells us that the influence of this factor cannot be great ; it mififht be calculated in the following wa\' : 162 S E R I V A L O R The losses by winding and throwing vary, as put forth in Chapter XIX between 0.2 and 2%. We will suppose in our case, a loss of 1%, the size being 13. In a thread of middling Regularity the thinnest parts will be of about size 7 ; these parts will chiefly break, and their weight being 1%, it resuhs : At the beginning a gram contained 092 meters ; of these were eliminated 0.01 gram = 13 meters (1 gram of size 7 containing about 1,300 meters) and there remained 0.99 grams = (092 — 13 =) 679 meters, which is equal to 686 meters to the gram. The alteration is then, in our case: 6 ^ = 0.9% 686 of which we might derive, as a general rule : The thickening of the sice, by elimitiation of the thinnest parts of the threads, is equal to the loss, in per cent. (B) When twisted the thread forms a screw-line which of course is no longer than the straight line. By following this longer way the thread becomes shorter and consequently of heavier size than it was before. In order to calculate this thickening of the size we must start from the diameter of the thread, for which purpose, however, the figures found in the last chapter must be somewhat altered. S E R I V A L O R 1G3 It is evident that the natural diameter of a soft body must become smaller under the pressure of the twisting. This diminution can be ascertained by a comparison between the diameters of the single and the twisted threads. If there had been no pressure, the latter would be (D =: single diameter) D J/^= 1.414D But it can be ascertained by the second method, explained in the last chapter, that it is only 1.21 D Consequently the diminution by pressure is 1/7, and in regard to the condition of the twisted thread X is equal to 12. (This condition resembles that within the mercury by which Professor Vignon found the specific weight of 0.9 to 1.1, viz., compressed but containing air. It was presumable, therefore, that a calculation of the specific weight on the basis of X=12 would give a similar result. In fact, it is 0.9.) My calculations of the shortening by twisting are therefore based. on: X = 12. When two threads are twisted together each spot of their .substance describes a curve which might be considered as the combination of two screw-lines. On the one hand each thread is turning around the com- 164 S E R 1 V A L O R mon axis — and this causes its shortening — on the other hand it is turning around its own axis, the consequence of which is a stowing, of which we shall speak later on. D being the diameter of the thread, P the progres- sion along the axis during one turn, L the length of the screw-line, then the shortening (S) : S = L — P and the thickening of size {T ) in per cent.: L — P T := P Consc(|rcntly we have for the twisting of a thread size 1 3 : (a) Tram, PiO turns to tiie meter: P = 8333 microns D = 45.6 microns L=y 83332 4- (45.G X 3.14)2 _ §334 S = l T = 0.012% (b) Organzine 425 turns to the meter : P = 2353 S = 1 T = 0.2% S E R I V A L O R 165 (c) Grenadine 1300 turns to the meter: P = 7T0 S = 13 T = 2% (d) Crepe, 3000 turns to the meter: P = 333 S = 30 T = 9% Adding to this the thickening by eHmination of the thinnest parts, the increase in size, in comparison to the original thread, is : For Tram ^% " Organzine 13^2% " Grenadine 3 % " Crepe 10 % These figures are ahered according to the losses in the winding. With threefold throwing the increase must be the same if the number of turns is rationally fixed. This number ought to be in inverse proportion to the diameter of the twisted thread, or rather to the square root of its weight — the latter being proportional to tlie diameter. Thus the number of turns (N) for threefold organzine of size 13 results from the ccjuation : N : |/"26"= 425 : ^SiT wliich gives N = 347. With this right method the greater shortening — 166 S E R I V A L O R owing to the longer way of the three threads — is bal- anced by the smaller number of turns. • In order to avoid the stowing of which we spoke before, the single thread receives a preliminary turn- ing in opposite direction of that of the twisting. But this well conceived expedient is turned into a disad- vantage if, as it is often the case now, the preliminary turns surpass those of the twisting. The stowing then occurs the opposite way, the organzine gets a granular aspect, becomes less lustrous and covers less well. Consequently the niunher of preliminary turns ought to be equal to that of the twisting. We shall now convert the above theoretical knowl- edge into such form as will enable anybody to make the accounts connected with throwing. We call: Theoretical size (Th. S.) the product of the multi- plication of the raw-size, with the number of threads to become united. Hence if we have to throw three threads of 11/13 (r= 12) the Th. S. is always 36, inde- pendent of the number of turns we give them. Of this Th. S. we must always draw the square- root ( ^ Th. S.). To this purpose table A has been made, which contains the \/ of all numbers from one to 200. If we divide the number of turns to the meter (T) by the j/ Th. S. the quotient we get forms the S E R I V A L O R 167 very distinctive feature of the throwing. We call this quotient the coefficient (Co.). Table B shows the shortenings of the thread corresponding to the differ- ence coefficients. Table C shows the coefficients for the usual kinds of throwings. We arrive from the number of turns to the meter to that one to the inch, dividing T with 40 {= — j consequently, multiplying the number of the turns to the inch with forty, we shall get T. (B) Coefficients of thr Giving {Co.) and shorten- ings in % of the Theoretical Si::e {Th. S.). Co.:... 138 190 237 271 308 341 368 393 417 440 462 483 503 %■ 1/2 1 V/2 2 2y2 3 3^ 44/2 551^ 6 6^, Co.: 522 540 558 576 594 612 630 647 664 680 696 %: 7 7i/> 8 81^ 9 9>^ 10 10^ 11 11^^ 12 (C) The usual throwings are the effects of the following Coefficients: Open, filling tram giving a soft touch Co. = 20 Hard tram, not filling giving a strong touch Co. = 30 Open, lustrous organzine for satins Co. =100 Medium organzine, less lustrous for armures Co. =120 Hard, lusteriess organzine for tafi^etas Co. =140 Lustrous grenadine ^ o. =:3()0 Lusteriess grenadine '..Co. =r2.>0 Lusterous crepe Co. =400 ]\redium crepe Co. =:500 Hard crepe Co. =600 o o H m 2 w H o CO W < a X ! C N C^ CC CC_ ■* Tj< tT •« «0 O O O l-_ J> !-_ C» 00 00 Ci O _ O C iH rH '>. CO CO re CO ro' cc co r^' c6 ro co co' ro ro ro cc o? co co co -^ ■*"■*' M< tj< tOt-GOC50r-IOJfCTj<»OOt-OOC50i-H(MCOTt00050 C t^i~-t^l^QOOOOOOOOOOOOOOOOOOOOiO!OiO:C5C50500>0>0 TH^T-HrHr-irH^THr-rHrti-lT-i-lr-i-HT-T- — ,-1-11-1-,-HCJ C3 ro t- iH o C5 ro t- T-H i-o C5 ro t^ T-i m 00 0} •-;: -* oc ci »ft oi ro |C OiCO_CO'*T^Tt.TtO«050J>l>000000©C; CC"— »— ^^(M >- W oi W M W oi N N ci CJ ci CJ N oi W CJ ci ci CO re co' r* co' n' CO _ i-HOJcoTfinot-oooso^cjro-^inot-oooCi— wco-ri0 i.O >-0 lO lO i.O O «5 cc o o «c «o O ^ O f^ i^ I'- l^ i> t^ , cot-iHOo>co5cooo?j«oo'nocot-(rj;c '*oo(Mi>r-(»n I C ci c i--; i-- cc oq C5 cs C O ^_ ^ ci w "> iH --I r-J IH T-i rt' r-i t-H ,-H r-' l-J i-H l-i i-i rH rt r-i r-I C^l M /rj" N CI N N i-(iHT-li-(,H^^iHi-lTHr-liHiHr-lr-(T-(T-irHTHi-li-i^r-(^^^,-( tOt-OOC50>HOJCO-*l.0 5Dl>OOOC— (J-lCO-^tOOt-OOOlO C coO».'5 0-^OlTj<05'+OOfOOOClt^WCCi-'0 tOO-^OO Ir; O'-;'-;M(MC0f0C0T»l-0000C;C~. ~0-H^ "> d s c d d d d d d d d d d d d d d d d c — ' —'-:—' ,h _ 'H5*-0 C t^_ l> 00 00_ 05 _ O i-H i-H *< Tt "O IC C£ u; l^ <» 00 C5 Ci ~> 00 00 00 00 oo' oi d ci d d d 05 05 o! d d ci d ci d d d d cJ d rH Ol^OOOOrHoiCOTfllOO^-OOCSC^-r'JCO'fl.OCSt^aOOSO C l^t~l--t-QOOOOOOOOOOOOCCCC»0005fflC5CiC>.0 ^ l^ ■* O Ci C-. <-0 i-H J~- CO c: -*i o -_ c! .-■: rt •-»■■>*< »o o o "> t-^ tJ t-' t,' i-J t,' t,; I-.' tJ l-J l^ t-.' t-^ 00 OC 00 OO «" OO' 00 00 00 00 00 oo" I _ ^IMCO'*<».OcOt-OOCsO.-llMCO-.*r30o«oo«oooo«ooocsi^i^t-i^i--i^ i_,CCC5O00t-«D>rtTt-; d CO -* •* o o I- J> 00 Oi o ^ moo lO >.o' in o o lo' o d d co d d w d d d d d d d t- t^ "> r^ 1-1 rH N CJ N W Cl CO CO CO CO CO co' co' -* Tj^' --t Tt Tl" -iji Tt<' Tji ■«!<' o 1— Ci CO •* 'O CC J^ OC C-. S rH S E R I V A L O R 169 First example : Question : What will be the size and the char- acter of the thread resulting from throwing 14 ends of size 10:55 with 1T0 turns to the meter (= 4% to the inch) ? Ansiver: (1) Th. S.= (10.55 X 14=) 147.7 den. (2) I/UtT = 12.15 (see table A). 170 (3) Co. = = 140 12.15 (4) Shortening (table B) of Co. 140 = ^% (5) Size = Th. S. plus ><% = 147.7+ 0.7 = 148.4 (6) Character: Hard organzine (table C). Second example : Question : How many turns to the meter and to the inch must I give to the size 9/11 = 10, three thread, to become a lusterless grenadine, and what will be the final size ? Ansii'er: (1) Th. S. = 4 X 10 = 30. (2) y^Q~= 5.48. (3) Co. for lusterless grenadine = 250. (4) T = 5.48 X 250 = 1370 turns to the meter = 34 to the inch. (5) Shortening = l->4%. (6) Final size = 30.5. 170 S E R I V A L O R Third example : Question : With what raw size must I begin, to arrive at "Poile" size 16, 2400 T (60 to the inch) ? Anszcer: (1) Th. S. = 16. (2) \/Ti\ = 4. 2400 (3) Co. = = 600 Counter- proof (4) Shortening for Co. 600 = 9%. (5) Raw size 16 minus 9fc ^ 14.56. (1) Th. S. = 14.56. (2) 1/1X56 = 3.8. 2400 (3) Co. = = 632 3.8 (4) Shortening of Co. 632 = 10%. (5) Final size = 14.56 + 10% = 16. (6) Character: Hard Crepe. Chapter XXII THE K I : i: D S IT M iGiiT be presumed that there is no mill in which the reeds are in perfect order, that is to sav. where all the dents are in ri^ht j)ro])ortion to their count. I am led to this opinion by the fact that while I was a manager I had great difliculty in procuring such reeds, seeing how little the reed-makers were used to keep strictly to instructions. And 3'et the right size of the dents is so important that without accuracy in this regard no reliable results may be reckoned upon. The friction of the silk against the dents increases with the number of threads, as the diminution of space produces a pressure, which, as proved by Coulombe, is proportional to the friction. Therefore it is not indifferent how much space is taken by the dents, and the proportion of their size to the clear space between them must be taken into con- sideration. 172 S E R I V A L O R TABLE OF PROPORTION BETWEEN DENTS IN THE CENTIMETER AND IN THE PARIS INCH. Dents in the Paris inch : 50 55 60 65 70 75 80 85 Dents in the centimeter: 18 20 22 24 26 28 30 Sl'^ Dents in the Paris inch: 90 95 100 105 110 115 120 Dents in the centimeter: 33 35 37 39 41 42^:; 44 Dents in the Paris inch : 125 130 135 140 145 150 155 Dents in the centimeter : 46 48 50 52 53!.4 55 57 Dents in the Paris inch: 160 165 170 175 Dents in the centimeter: 59 61 63 65 The "No." expresses the number of dents in the total size of the % Paris inch (= 67^ Ctm.). Conse- quently the size of one dent is : No.: 35 36 37 38 39 40 41 42 43 44 Microns: 193 187 182 178 173 169 165 161 157 153 No.: 45 46 47 48 49 50 51 52 53 54 Microns : 1.50 147 144 141 138 135 132 129 127 125 No.: 55 56 57 58 59 (JO 01 6:2 63 64 Microns : 122 121 119 117 115 113 111 109 107 105 No.: 65 66 67 68 69 70 71 72 73 74 Microns : 104 102 101 99 98 97 95 94 92 91 No.: 75 76 77 78 79 80 81 82 83 84 Microns : 90 89 88 87 86 84 83 82 81 80 No.: 85 86 87 88 89 90 91 92 93 94 Microns : 79 78^^ 78 77 76 75 74 73 72^ 1 72 No.: 95 96 97 98 99 1 00 Microns: 71 701/ '2 70 69 68'/ ; 68 SERIVALOR n3 By choosing, in this table, the "No." in proportion to the number of dents, the reeds might be so con- structed that they all have the same clear space, which ought to be in the neighborhood of 0.7 Ctm. But it is not quite easy to have such reeds made by the reed-makers who are accustomed to employ cer- tain wires for certain "Xo." thus producing reeds whose clear space varies from 4 to (i.^i millimeters to the centimeter. They are led by no distinct principle in this regard, but simply by the fact that it is more convenient for them to fix thicker dents with thinner wires than vice versa. If a reed with forty dents No. 00 to the centimeter is ordered of them, they generally will deliver dents No. 70, trusting that they cannot be controlled. This, however, is quite easy with the aid of our table, and of a slide-gauge indicating 0.1 millimeter, if one orders a little bit more of length than wanted and takes off 5-10 dents, measuring their total size ; or with the help of a microscope, without taking ofif any dents. If one protests against the "No." not in accord- ance with that ordered they will reply that dents No. 90 are not durable enough, and on the question why dents no should be less durable in a reed with 90 dents than in one with 50 to the centimeter, they will offer other excuses, declaring finally : 'Tt is impossible." It is a fact that fine reeds are less lasting, but this will keep no manufacturer from employing them, see- ing that they furnish a finer tissue. It appears, more- m S E R I V A L O R TABLE OF CLEAR SPACES Dents per ctm. : IS 19 20 21 22 23 24 25 m/m Clear space, : 70;} 700 697 694 691 G88 685 682 100 No. of dents: 41 43 45 47 48 50 51 53 Dents per ctm.: 26 27 28 29 30 31 32 33 m/m Clear space : 679 676 673 670 667 664 661 658 100 No. of dents: 55 57 58 60 61 62 63 65 Dents per ctm. : 34 35 36 37 38 39 40 41 m/m Clear space : 655 652 649 646 643 640 637 634 100 No. of dents: 66 68 69 71 72 73 74 76 Dents per ctm.: 42 43 44 45 46 47 48 49 m/m Clear space ■ : 6;!1 628 625 622 619 616 613 610 100 No. of dents: 77 78 79 SO 81 82 83 84 Dents per ctm.: 50 51 52 53 54 55 56 57 m/m Clear space : 607 604 601 .598 .595 592 589 586 100 No. of dents : 85 86 88 89 90 91 92 93 Dents per ctm.: 58 59 60 61 62 63 64 m/m Clear space : 583 580 577 574 571 568 565 100 No. of dents: 94 95 96 97 98 99 100 S E R I V A L O R 175 over, that the making of fine reeds with a clear space of 0.7 Ctm. offers some difficuhies. Until these are overcome we must confine ourselves to calculating a series whose clear spaces diminish, with the fineness, from 0.7 to 0.565 Ctm. Comparing the space demanded by the respective warp (see Chapter XX) to that left to it, we find out the limit where the w-arp-threads begin to be pressed together and therefore are exposed to increased fric- tion. From this moment the difficulties increase rapidly and the quality of silk must get proportionally better, if weaving is to be at all possible. On those circumstances is based the table of Chap- ter VJIT (cohesion), the result of three years' studies and work. Chapter XXIil soMi HINTS ABOL r POWER \vj;a\ing XTo SILK is good enough if the loom is not in perfect ■^ ^ order. While I was still at the inception of my studies the results obtained were sometimes apparently con- tradicted by experience. A bale of silk which had been classified as good under test, jiresented in the weaving unexpected difficulties which were sometimes hard to overcome. But in ninety-nine out of one hundred cases they lay in the loom itself, and by finding their origin and trying to avoid them thereafter, I was led to establish the following rules which are not to be found in weav- in