respective diameters of the wires. Nor did the wires hardened on any particular day show like thermo-electric qualities. It is conceivable, for instance, where currents of different intensity are employed, that the resulting difference in the degrees of red heat imparted to the wires would find its expression in a correspondingly marked difference in the degrees of hardness. The true cause of the difference of character observed must therefore be ascribed to chemical composition. In other words, the groups are distinguishable one from another by the respective amounts of carbon contained in the wires. The rods were not all received at the same time. Without doubt the maximum of hardness attainable by sudden chilling is essentially conditioned by the degree of carburation of the steel rod, and to a very small degree only by the impurities present. In a general way we may state that the maximum in question is a characteristic datum for the type of steel under experiment. In Chapter VII of the present memoir we shall have occasion to discuss this subject in detail.

Maximum hardness reached.-The greatest hardness met with in this work was that possessed by two rods of the diameters 0.056 cm. and 0.073 cm., respectively. The thermo-electric constant a here reached the exceptionally small value a= — 2·60 (microvolts), and the specific resistance, at ordinary temperatures, was as large as 45 microhms, cm cm2. Unfortunately most of these very hard wires had to be dis carded, because in the earlier experiments boiling water had been used in investigating their thermo-electric powers. We subsequently found that 100 produces a very pronounced annealing effect. In the later experiments with glass hard wires water of only 400 was employed, and this but for a very short period of time.

Temperature of ignition.-Jarolimek and Ackermann27 in their experiments on steel arrived at the important result that the rapidity of the first part of the chilling of red-hot steel, say from 6000 or 7000 to 3000 or 400°, is far more essential as regards the degree of hardness obtained than the further cooling from 300°-400° to zero. It is easily possible to harden a rod very perceptibly by cooling it from bright redness in a metallic bath (Zn, Pb) at 400°. Such a process combines in one the operations of chilling and of annealing. Cooling from 3000 or 400°, however, produces no effect. Again, Chernoff had previously found that if the temperature from which steel is chilled be supposed to increase continuously, no observable effect will be apparent until a temperature in dark cherry-red heat is reached, when glass-hardness is suddenly attained.28 Our experience is in perfect accord with these results. The phenomenon was strikingly manifested both in rods of the same diameter and in

27 Jarolimek u. Ackermann, Zeitschr. für das chem. Grossgewerbe, 1880. Similar results, we believe, were published by the distinguished American engineer, Mr. Joshua Rose, but we have been unable to find them.

28 D. Chernoff: Vortrag gehalten in der russischen Technischen Gesellschaft, im April u. Mai 1868.

those of different diameters. In the first instance we may cite our observations with comparatively thick steel rods, which the available current was able to heat to dark redness only. After chilling these remained soft and pliable for more than one-third of their length; but at a particular point of the wire its mechanical condition changed suddenly to brittle hardness, and this despite the fact that during the heating a change of intensity of redness along the parts of the wire in question was scarcely discernible. In the second instance we frequently noted that where the intensity of current used was sufficient to impart hardness readily to a given comparatively thick class of wires, this was no longer possible for the next larger dimension, notwithstanding the fact that the respective thicknesses varied as little as 0.125 cm. and 0.145 cm. It is furthermore to be added that the sudden change in question is equally apparent both in the mechanical as well as in the electrical properties of steel.

The experiments made would certainly not warrant the forced assumption that the phenomenon in question partakes of the nature of a true discontinuity. Some molecular change occurring at an extremely rapid rate is alone to be understood. In a general way, however, it may be stated that a certain critical temperature in red heat must be surpassed if sudden cooling is to produce glass-hardness; otherwise the steel remains soft.

a

Thermo-electric maxima and neutral points.-In this place a final remark may be added. During the experiments we had frequent occasion to observe "neutral points" (occurring at the mean temperature ¿(T+t)=1% of the junctions of the thermo-element, the temperature at which the electromotive force passes with a change of sign through zero) of comparatively low value. Many of the hard rods differed but slightly from silver as regards their thermo-electric properties. Maxima of electromotive force at low temperatures were even more frequently obtained. A number of examples of this kind will be found in the tables below.

BEHAVIOR OF HARD STEEL RODS ANNEALED IN HOT OIL BATHS.

W

Manipulation-Having thus in hand an assortment of glass-hard rods of excellent quality, it was our next endeavor to reduce the degrees of hardness of these consecutively by equal amounts. In other the problem was so to anneal the rods that between the glasshe soft states as extremes a great number of intermediate iy equidistant states might be obtained. We commenced ons with this end in view by heating the hard steel in a ch of linseed oil very gradually to different temperatures, re

moving samples of the series of immersed rods at different stages of the process. Inequalities of temperature in the bath were reduced to a minimum by constant stirring. We also adopted an inverse method

of procedure, in which, when a desired temperature was reached, the hard wires selected were submerged in the oil on a false bottom of wire gauze, and the whole allowed to cool. After several days they were examined both as to their thermo-electric power and their specific resistance.

Results. The results of these measurements are given in the following table. It will be remembered that e (observed or calculated as specified) is the electromotive force in microvolts for the temperature T and t (centigrade) of the junctions. a and b are the thermo-electric to microhm the specific resistance at

constants of Avenarius, 8, (cm2t

the temperature t, p (cm.) finally the radius of the wires:

TABLE 7.-Thermo-electrics and conductivity of steel wires, annealed in oil baths.

TABLE 7.-Thermo-electrics and conductivity of steel wires, annealed in oil baths-Cont'd.

Digest.—If we arrange these different degrees of hardness, expressed both thermo-electrically and in terms of their specific resistance, with reference to continuous variation of the former quantity, we obtain

the following perspicuous tabular comparison. The table shows that our endeavor to reach a great number of symmetrically distributed degrees of hardness systematically, was only partially successful.

TABLE 8.-Thermo-electric position and conductivity of annealed steel.

ON THE BEARING OF THE TIME OF EXPOSURE ON THE EFFICACY

OF ANNEALING.

Low annealing temperatures.-In the foregoing experiments the lowest temperature employed was 1:00. It will be seen that the annealing effect thus produced is strikingly large. This observation naturally suggested the question as to what results are to be expected when the annealing is conducted even at lower temperatures. The inquiry would have an immediate practical bearing: It will be remembered that in the thermo-electric measurements it is desirable to raise one of the junctions to a high temperature relatively to the other. We are led to ask, therefore, how high this temperature may be chosen without destroying uniformity of temper and producing partial annealing at one end of the rod.

Results for 1000.-We began the preliminary experiments with the two rods, Nos. 24 and 25, of nearly the same thicknesses, 0.0574 cm. and 0.0554 cm., respectively, but of different degrees of glass-hardness. Thermo-electric measurements only were made, with results for the glass-hard state as follows:

TABLE 9.-Thermo-electric power of glass-hard wires.