Consequently their number must be constantly varying in accordance with their condition. These facts show that the difficulty of maintaining regularity in a thread is very great. Nevertheless, this regularity is one of the principal factors of the value of a thread of "grege," and this to such an extent that badly reeled silks are sold at from twenty to twenty-five francs a kilogramme less than those which are satisfactorily regular.

The difficulty of this hand labor can be still better understood if it be remembered that the reeler being obliged to watch at every moment the unwinding of each cocoon, in order to obtain one pound of well reeled silk, she must incessantly watch, and without a moment of distraction, the unwinding of about two thousand seven hundred miles of silk filaments. For nine pounds of silk, she reels a length of filament sufficient to girdle the earth. The manufacturer, therefore, cannot and must not depend only on the constant attention that each reeler should give to the work confided to her care. He is obliged to have overseers who constantly watch the reelers, so that the defects in the work of any single reeler, who otherwise might not give the attention required by her work, will not greatly diminish the value either of her own work or that of several other reelers whose silk is often combined to form a single lot. In addition to the ordinary hand labor, considerable expense is thus necessitated for the watching of the reelers.

Enough has now been said, we think, to give a good idea of silk reeling, as usually practiced, and to show how much it is behind other textile arts from a mechanical point of view. To any one at all familiar with industrial work, or possessing the least power of analysis or calculation, it is evident that a process carried on in so primitive a manner is entirely unsuitable for use in any country in which the conditions of labor are such as to demand its most advantageous employment. In the United States, for instance, or in England, silk reeling, as a great national industry, would be out of the question unless more mechanical means for doing it could be devised. The English climate is not suitable for the raising of cocoons, and in consequence the matter has not attracted very much attention in this country. But America is very differently situated. Previous to 1876 it had been abundantly demonstrated that cocoons could be raised to great advantage in many parts of that country. The only question was whether they could be reeled. In fact, it was stated at the time that the question of reeling silk presented a striking analogy to the question of cotton before the invention of the "gin." It will be remembered that cotton raising was several times tried in the United States, and abandoned because the fiber could not be profitably prepared for the market. The impossibility of competing with India and other cheap labor countries in this work became at least a fact fully demonstrated, and any hope that cotton would ever be produced in America was confined to the breasts of a few enthusiasts.

As soon, however, as it was shown that the machine invented by Eli Whitney would make it possible to do this work mechanically, the conditions were changed; cotton raising become not only possible, but the staple industry of a great part of the country; the population was rapidly increased, the value of real estate multiplied, and within a comparatively short time the United States became the leading cotton country of the world. For many years much more cotton has been grown in America than in all the other countries of the world combined; and it is interesting to note that both the immense agricultural wealth of America and the supply required for the cotton industry of England flow directly from the invention of the cotton gin.

Attention was turned in 1876 to silk raising, and it was found that all the conditions for producing cocoons of good quality and at low cost were most favorable. It was, however, useless to raise cocoons unless they could be utilized; in a word, it was seen that the country needed silk-reeling machinery in 1876, as it had needed cotton-ginning machinery in 1790. Under these conditions, Mr. Edward W. Serrell, Jr., an engineer of New York, undertook the study of the matter, and soon became convinced that the production of such machinery was feasible. He devoted his time to this work, and by 1880 had pushed his investigations as far as was possible in a country where silk reeling was not commercially carried on. He then went to France, where he has since been incessantly engaged in the heart of the silk-reeling district in perfecting, reducing to practice, and applying his improvements and inventions. The success obtained was such that Mr. Serrell has been enabled to interest many of the principal silk producers of the Continent in his work, and a revolution in silk reeling is being gradually brought about, for, strangely enough, he found that the work which he had undertaken solely for America was of equal importance for all silk-producing countries.

We have described the processes by which cocoons are ordinarily cooked and brushed, these being the first processes of the filature. Instead of first softening the gum of the cocoons and then attacking the floss with the points of a brush, Mr. Serrell places the cocoons in a receptacle full of boiling water, in which by various means violent reciprocating or vortex currents are produced. The result is that by the action of the water itself and the rubbing of the cocoons one against the other the floss is removed, carrying with it the end of the continuous filament without unduly softening the cocoon or exposing any of the more delicate filament to the rough action of the brush, as has hitherto been the case. The advantages of this process will be readily understood. In brushing after the ordinary manner, the point of the brush is almost sure to come into contact with and to break some of the filament forming the body of the cocoon. When this occurs, and the cocoon is sent to be reeled, it naturally becomes detached when the unwinding reaches the point at which the break exists. It then has to be sent back, and the end of the filament detached by brushing over again, when several layers of filament are inevitably caught by the brush and wasted, and very probably some other part of the filament is cut. This accounts for the enormous waste which occurs in silk reeling, and to which we have referred. Its importance will be appreciated when it is remembered that every pound of fiber thus dragged off by the brush represents a net loss of about 19s. at the present low prices.

The mechanical details by which Mr. Serrell carries out this process vary somewhat according to the nature of the different cocoons to be treated. In one type of machine the water is caused to surge in and out of a metal vessel with perforated sides; in another a vertical brush is rapidly raised and lowered, agitating the water in a basin, without, however, actually touching the cocoons. After a certain number of strokes the brush is automatically raised, when the ends of the filaments are found to adhere to it, having been swept against it by the scouring action of the water. The cleaning of the cocoons is performed by means of a mechanism also entirely new. In the brushing machinery the floss is loosened and partially detached from the cocoon. The object of the cleaning machine is to thoroughly complete the operation. To this end the cocoons are floated under a plate, and the floss passed up through a slot in the latter. A rapid to and fro horizontal movement is given to the plate, and those cocoons from which the floss has been entirely removed easily give off a few inches of their filament, and allow themselves to be pushed on one side, which is accomplished by the cocoons which still have some floss adhering to them; because these latter, not being free to pay off, are drawn up to the slot in the plate, and by its motion are rapidly washed backward and forward in the water. This washing soon causes all the cocoons to be