furnaces may be taken down in a few minutes. They are capable of reducing the maximum variation of temperature along 8 in. of the central portion of the 18-in. tube from 30°C. to less than 10°C. within the range 200° to 1,000°, using a single adjustment.

In all of the annealing work, the test pieces were placed in a hot furnace and rapidly brought to the annealing temperature which was then kept constant for the desired period by adjusting the external resistance. With the present set, a preheating period of 10 min, is to be added to the annealing period of 30 min. The results are shown in Table I.

TABLE I.-Early Effects of Annealing upon Physical Properties of Worked Brass

The two sets of numerical values are consistent and seem to indicate that there is a slight but unmistakable increase in tensile strength and a corresponding decrease in elongation produced by heating at 200°C. It would seem that Bengough and Hudson's curves indicate more accurately than Grard's the true conditions brought about by heating at low temperatures.

The precise nature of the changes which take place in worked metal as a result of exposure to very moderate temperatures, or even spontaneously, is not understood. Any attempt to study these changes with the aid of the microscope is certain to fail, since the accompanying structural alterations are too minute or too indefinite to be followed. In the present case, careful comparisons between the two sets of strips were made with the aid of a 12-in. immersion objective without detecting any difference. Both the heated and the original strips show the typical folds, or wrinkles, which are prominent in heavily worked brass and may be seen under low power in the photomicrographs of Plate II, or under high power (X 1,000) in Fig. A, Plate I, leaving out the central, recrystallized portion.

It is obvious that since no alteration of structure can be associated with these effects they must be of a mechanical nature, or they must be due to some reorganization of the particles which does not suffice to change the etching characteristics of the metal. The former interpretation is practically substantiated by the recent work of E. Heyns in which the distribution of internal strains in cold-worked metal before and after heat treatment is determined by means of special tests. This subject is handled at length and in a masterly manner by Heyn. We will merely draw attention to his observation that reheating at temperatures even below the softening point is productive of "remarkable effects toward diminishing the degree of internal strains" with the comment that similar effects in the present experiments are in all probability responsible for the differences observed in the ordinary tensile test. Heating at or above the earliest softening point would entirely remove internal strain but would also lower the tensile strength. It would seem from the experiments in hand that the frequent destructive effect of aging (seasoncracking) in severely worked metal, which is undoubtedly due to relief of internal strain from some local cause, could be avoided in every case by suitable heat treatment without loss of tensile strength. The present tests show that there may actually be an increase in tensile strength associated with the partial release of strains in some rolled shapes. Time Required at a Number of Different Temperatures between 225° and 325°C. to Produce a Drop of 3 Points in Scleroscopic Hardness

in the Case of 70/30 Brass which had Received 40 Per
Cent. Reduction by Rolling after a Dead Soft

Anneal in the Mill

The same material was used in these tests as in those just described. A number of 1-in. pieces cut from the 8-in. strips were placed in the furnace at the given temperature and removed at intervals.

TABLE II. Factors Corresponding to a Drop of 3 Points in Scleroscopic Hardness after 40 Per Cent. Reduction by Cold-Rolling

3 E. Heyn: Internal Strains in Cold-Wrought Metals and Some Troubles Caused Thereby, Journal of the Institute of Metals, vol. xii, pp. 3–37 (1914).

At 425° and 375°, the strip pieces could not be brought to the furnace temperature, which required about 5 min., without losing more than the desired 3 points on the scleroscope scale. In the former case, the loss was 13 points, and in the latter, 8 points. Between 325° and 225°, the time necessary for a 3-point drop could be measured and the values obtained are given in Table II.

Fig. 1 shows the above results in graphical form. From a practical standpoint, if we are to speak of a "critical" temperature at which softening will begin in any particular mixture which has been given a stated. reduction by cold-working, we must not omit certain qualifications, since the same effect will be produced at a number of different temperatures depending upon the time allowed. From the indications of this curve and from other considerations to be shown later, there is reason to believe

Time necessary to cause drop of 3 points in scleroscopic hardness. FIG. 1.-CARTridge Brass (70/30) Reduced 40 PER CENT. IN AREA OF SECTION BY COLD-ROLLING.

that there is a minimum temperature corresponding to a given degree of reduction at which the first softening effects may occur, but at this temperature the change proceeds very slowly, so that a period of days, or months, may be required to bring about equilibrium. The particular equilibrium involved will be discussed in another section of this paper.

In the present case, the arbitrarily chosen unit of measurable softening* is realized at a temperature as low as 225°, but only after a period of about 121⁄2 days, while, at a temperature only 50° higher, the same effect is produced in about 5 hr., and at a temperature still higher by 50°,

*We have chosen 3 points on the scleroscope scale as the measure of appreciable softening since there can be no doubt from an experimental standpoint that the metal has softened when this indication is obtained. A drop of two points is not as easy to detect, while an indicated drop of 1 point is always open to doubt.

viz., at 325°, the effect is produced in a period of minutes. None of the values shown on the curve are equilibrium values, since further softening occurs at any given temperature on longer exposure; from the flattening of the curve, however, it is apparent that the lowest value given, viz., 225° is not far above the true equilibrium temperature for this degree of softening.

Bearing these facts in mind, it is clear that any set of curves giving the physical properties of a commercial mixture after anneal for a fixed period of time following a given reduction by cold-working requires that the stated period of anneal be rigidly adhered to within the critical range of incipient effects, in order that the results of test may be properly duplicated. At higher temperatures, when all changes are rapid, approximate equilibrium values are reached in comparatively brief periods of time and some latitude in this respect may be allowed without materially affecting the results. For example, in the case of cartridge brass (70/30±3), which has received a heavy reduction, at all temperatures above 325°C., variation of the annealing period between 2 hr. and 1 hr. will not bring about material change in properties; at somewhat lower temperatures, however, incipient effects are encountered and the time factor must be carefully established by test.

Early Stages of Recrystallization in Material which has Received a High Degree of Reduction by Cold Work

Rose observes that the change in hardness and the recrystallization, in the case of gold, take place almost simultaneously, with a slight lag in the visible recrystallization; especially if the specimen is heated for a short time only. As far as can be inferred, his observations were all made at low magnification and, in case of a slight drop in hardness, he failed to detect recrystallization. As a result of the experiments just described, we are in possession of a large number of small pieces of metal which show different degrees of hardness varying from 33, that of the original cold-rolled metal, to 14, and these variations were brought about by heating for different periods of time at the various temperatures within the range, 225° to 325°. A sufficient number of these were carefully. examined to prove beyond doubt that any positive indication of softening by the scleroscope can be detected in the form of recrystallization under the microscope. This applies only to material which has received comparatively heavy reductions, as will be pointed out more clearly in a later section.

In order to detect recrystallization of this order, microscopic examination under high-power is necessary, as the first visible elements of recrystallization are extremely minute. Before describing these effects,

2 Journal of the Institute of Metals, vol. viii, pp. 86-125 (1912).

it may be well to refer briefly to the ordinary appearance of brass after different degrees of strain-hardening, as seen under the moderate magnifications usually adopted in brass work. A set of photomicrographs* representing the effect of cold-rolling in stages from gages 18 to 25 (B. & S.) using alpha brass containing about 64 per cent. copper, 0.2 per cent. lead and 0.04 per cent. iron is shown in Plate II. The reductions are given with the micrographs. The samples were etched with ammonia and hydrogen peroxide: The common high brass, from which these micrographs were made, and cartridge metal, which is the subject of the present. experiments, both show the same deformational characteristics. It may be noted that Fig. F, Plate II represents approximately the same degree of reduction as that received by the metal used for the present tests.

In general, reductions of 5 or 10 per cent. have little or no effect on the etching properties of alpha brass. Photographs taken before and after deformation are very similar; practically the only alteration due to moderate working of the metal which may be observed in the photograph is a slight displacement of the normally rectilinear boundaries between parts of a twinned grain, so that they frequently appear curved. Some of these curved bands may be seen in Fig. B, but, more prominently, in the succeeding figures. When the reduction reaches about 20 per cent., the etching characteristics of the metal begin to change; there is a loss of contrast between differently orientated grains owing to the development of a secondary structure within the grains themselves. This is barely discernible in Fig. C, but becomes more prominent as we pass to higher reductions in the succeeding figures. At a reduction of about 30 per cent., as in Fig. D, the direction of extension in cold-working becomes apparent, in that the grains themselves show a prevailing direction of elongation.

The intergranular secondary structure due to deformation consists in the main of groups of curved or wavy lines which possess the same general direction in each grain, or homogeneous portion of a twinned grain, and increase in number and prominence with the degree of reduction. They incline to a direction at right angles to the direction of elongation and it appears that, in the squeezing of the grains between the rolls, the most destructive movements affecting the integrity and homogeneity of the grain substance have occurred chiefly along the gliding planes which, in any given grain, are nearest at right angles to the direction of elongation.

In the absence of any generally accepted term of definition, these lines may be called, simply, lines of deformation. They are the most prominent and the most characteristic lines which are produced by any deformational process. In some laboratories, the term slip bands has been dis

These photomicrographs were prepared by Philip Davidson in the Hammond Laboratory for the Scovill Manufacturing Co. who have kindly permitted us to use them in the present connection.