Numerous tests were carried out at the works of the author under his personal supervision by Mr. J. M. Gilman, who had previous experience in testing under Dean Turneaure of the Engineering Department of the University of Wisconsin. Every available convenience required for investigation was utilized. The testing machine was largely "home-made," with the exception of the excellent hydraulic apparatus which was generously loaned by Mr. F. H. Stillman of the firm of Watson-Stillman Co., manufacturers of hydraulic appliances, New York, to whom the author desires to record his obligations and thanks.

The operation of the testing machine required two observers, also a recorder and two handy-men, and the use of a steam derrick for moving the beams. When making a test, the beam was placed in position, with the observers stationed on either side to read the two sets of Professor Johnson's extensometers and the deflectometer. The observers announced the action of the beam to the recorder who entered the same on the prepared blanks. One handy-man operated the hydraulic force pump and another applied the weights as required on the spinning accumulator gauge. The application of load could thus be made at any rate of speed or repeated as often as needed. This feature of a hydraulic apparatus is especially useful for repetitive tests and for obtaining permanent set and the effect of speed in loading.

The spinning accumulator gauge was calibrated, and checked with a dial gauge for pressures up to 1,000 pounds per square inch. For pressure above 1,000 pounds and up to 5,000 pounds per square inch, the maximum capacity of the machine, the spinning accumulator gauge was used alone. A light oil instead of water was used for the hydraulic apparatus to avoid danger from freezing. On the completion of a test the handy-man replaced the tested specimen with an untested one. It required about one-half day to make a complete test of each specimen and to place another in position; usually two tests were completed in one day.

In calibrating the machine to get the actual applied load, two weighing machines having platform scales were placed on either side of the apparatus and supported the ends of a beam to take the weight at points of application of the load. This beam transmits the total load received to both scales which were read and recorded. Tests were made at several points of application, as

shown in photograph illustrating the calibrating of the testing machine. Some of these were with 50-pound increments per half minute, and others with 100-pound increments at 1 and 2 minute intervals. These loadings were applied up to 5,000 pounds

per square inch, which was the limit of the machine. The actual

weight shown by the scales at each interval was recorded. Upon reaching the limit of the machine the load was released in a manner similar to its application.

It was desired to discover whether there was any material difference in applying the load at differing rates of speed. As a result it was determined that increments of 100 pounds per minute were most convenient, and secured the most uniform results. The records were plotted and the curves show that in every test, upon running up the load, the actual exceeded the computed weight, and in releasing this condition was reversed; the difference was probably due to internal friction of the hydraulic jacks.

When the load was retained at 5,000 pounds per square inch for 8 minutes there was a falling off of pressure.

In testing specimens the webs of the I beams finally buckled at maximum loads due to stretching of the spring steel fulcrums. To guard against a repetition of this we now use double I beams. The results of our tests are available in any really scientific research work.

The accompanying drawings show the testing machine in complete detail and make a lengthy description unnecessary.

SPECIFICATIONS FOR COTTON TAPES

FOR ELECTRICAL PURPOSES.

By R. D. DEWOLF.

In the making of a specification for a manufactured product, the more complex the manufacturing process, the more difficult it becomes to differentiate between the necessary and the superfluous restrictions to be embodied in the specification. Dr. Dudley has said, "A specification should contain the fewest restrictions consistent with obtaining the material desired," and "the service which the material is to perform should determine the limitations of a specification." These considerations are quite generally overlooked in the case of woven fabrics, the consumer binding the manufacturer down to very narrow limits in his selection of raw material, method of production, etc.

In making specifications for woven fabrics, especially in the cases of tapes, there is a general tendency to select a fabric which has been found to be satisfactory for the purpose in view, to analyze this, and to embody this analysis in permanent form as a specification. This method is open to criticism in that it allows the manufacturer no leeway in his selection of yarns, combinations of same, or method of manufacture. Specifications have recently been issued by the Westinghouse Electric and Manufacturing Company in which an endeavor has been made to overcome these objections, and the specifications so drawn as to allow the manufacturer the greatest possible variation consistent with the production of a fabric possessing the required characteristics.

Among the desirable qualities which should be embodied in a tape suitable for electrical purposes, we may note:

1. Sufficient strength to withstand any strain likely to be placed upon it.

2. A sufficiently close weave, varying with the thickness of the tape, to give a neat appearance and withstand abrasion.