spindle, on the theory that the spindle acted like a top, and found its centre of rotation under an unbalanced load. This theory has since been discarded by experts, it now being thought that the advantages of the Rabbeth spindle are derived, first, from the cushioning effect of the loose bearing; and second, from the additional cushioning effect of the packing interposed between the bolster bearing and the surrounding case, both taken in connection with a sleeve whirl surrounding the tube containing the bearings. The spindle does not centre itself, but runs out of centre with less jar and vibration and heat, and thus is enabled to bear a greatly increased speed, and to run with less power.

About the time of the invention of the Rabbeth spindle other inventors produced different structures in a similar line. John Birkenhead of Mansfield, Mass., in fact preceded Rabbeth in making a structure having a yielding bolster in combination with a sleeve whirl spindle.

BIRKENHEAD SPINDLE.

The Birkenhead yielding bolster, however, was attached to a fixed step, and the opportunity for yielding was given by an elastic connection between the bolster and step. This type of spindle was somewhat used, and it went into considerable use in a form known as the "Eureka" spindle, patented by Chapman some years since. In that form the bolster was held rigidly, and the step allowed to yield by means of elastic connections between them. Experience has shown that this construction is less advantageous than a construction that has freedom at both bearings.

A little after Rabbeth's invention, Mr. John E. Atwood of Stonington, Conn., made a sleeve whirl spindle in which the bolster and step were made in one piece, and attached "in a yielding manner" to the surrounding shell or bolster case. This structure has gone into use to an extent of hundreds of thousands in silk spinning, but not extensively in cotton spinning, because of the great liability of the spindle to stand out of the centre of the ring at some points of the traverse of the ring rail. In silk spinning there is no ring or ring rail, the process being entirely different from that necessary in spinning cotton. The silk is spun off the spindle and the cotton is wound upon it.

At the time of the Rabbeth experiments, Mr. Albert R. Sherman was employed as draftsman and engineer in the shop of Messrs. Fales & Jenks, where most of these experiments were tried. He was of opinion that the packing between the bolster and supporting tube of the Rabbeth spindle was objectionable, and liable to need frequent replacement.

As early as the fall of 1878 he made spindle structures in which the packing was omitted. The bolster and step were made in one piece, and loosely fitted, while the supporting tube, unlike the supporting tube of Atwood's, was closed at the bottom and formed an oil reservoir. Unlike both Rabbeth and Atwood, he used no packing or spring connection between the bearings and the case that held them, but the oil in the supporting tube was supposed to act as a cushion between the combined bolster and step and the surrounding case during the vibrations of the spindle.

Opposite the band pull, the outer surface of the bolster was made cylindrical, and the inside of the supporting tube in the same plane was made cylindrical also, though of a slightly larger diameter, so that the band pull would not tend to deflect the spindle from the perpendicular in running.

Mr. Sherman, also by a positive lock or connection prevented the revolution of his bolster with the spindl The Sherman experiments continued until the lat

of 1880, when an entire frame was put in operation. From that time the sale was rapid, and the Sherman spindle is one of the standard spindles of to-day, several hundred thousand

of them being in use. The spindle is clearly within the Rabbeth principle, though it differs from the Rabbeth spindle, as known in the trade, in several details of construction. Another experimenter in sleeve whirl spindles, closely following Rabbeth in the year 1878 or early in 1879, was Mr. Thomas Mayor of Providence, R. I. Mr. Mayor had previously been an inventor in both the fly and ring frame spindles, and it was claimed that he had used an elevated bolster on ring frames prior to Mr. Sawyer's invention. His idea on the sleeve whirl spindle was a very ingenious one, and capable of practical use, though it has not been adopted to any great extent, because the owners of his patents were also the owners of the Rabbeth spindle, and preferred the latter. He made a combined bolster and step, having a swell on its outer surface, which fitted into a socket in the supporting tube, forming a kind of ball and socket joint. The band groove of the sleeve whirl surrounding the supporting tube and its bearings was located substantially in the horizontal plane of this ball, that the pull of the band had little or no effect upon the perpendicular position of the spindle. In case of an unbalanced load upon the spindle, the structure allowed it to gyrate with this ball as a fulcrum. A little later, Mr. Mayor

MAYOR SPindle.

placed springs or packing opposite the upper or lower ends of this combined bolster and step, so that, during gyration, there was a constant tendency to press the spindle back to a perpendicular position.

There have been several patented modifications of the Mayor principle, but, like the original, none of them have been adopted to any great extent. Mr. C. D. McDuffie, of the Manchester Mills, has been recently experimenting in this direction, and reports considerable success. As before stated, the Rabbeth top spindle was a great step in advance in the art of spinning, even over Sawyer's.

The Sawyer spindle was limited in speed. With an unbalanced load it would vibrate and gyrate, at more than seven thousand five hundred turns per minute, so as to become useless. The Rabbeth spindle, on the contrary, will bear any speed desired, and the limit of production of the frame is transferred from the speed that the spindles will bear, to the speed with which operatives can make good piecings of yarn broken in the operation of spinning. And I might add, on fine numbers the difficulty with which travellers are attached to the rings, and with the very finest numbers the difficulty in getting a traveller of the desired weight and shape, so that at the Willimantic Woollen Mills they produce yarns at the highest speed at which they run. From nine to ten thousand revolutions per minute is the speed at which they are customarily run on medium yarns.

The power required to drive them at a speed of nine thousand does not exceed the power required to drive the common spindle at a speed of five thousand five hundred.

Returning to Rabbeth's early experiments, the first few spindles tested in the shop of Messrs. Fales & Jenks attracted more attention, and deservedly, than any other invention in cotton manufacturing during the present generation. The spindles were found capable of unlimited speed, and capable of carrying absurdly unbalanced loads, one of the tests made for visitors being the running of a broom