Chapter LXVII: Part I

"To illustrate the practical importance of knowing the effects of reiterated heating and cooling on iron plates, one of the most obvious examples is the action of heat upon the plates of boilers which are alternately heated and cooled, as in use or otherwise. When in use, the plates above the fire are subjected to the fierce flame of the furnace on one side, and on the other side to a temperature approximating to that of the steam and water in the boiler. Where the conducting surfaces of the metal are thickened at the riveted seams, a source of danger is frequently revealed in the appearance of what are known as 'seam-rips.'

"The long egg-ended boilers, much used in the North of England, are very subject to this breaking away of the seams. From some tests made by the writer on iron cut from the plates of two different boilers which had ripped at the seams, and one of which seam-rips had led to an explosion resulting in the destruction of much property, though happily of no lives, it was found that the heat acting on the bottom of the boiler had, through time, so affected the iron at the seam as to make it brittle, apparently crystalline in fracture, and of small tensile strength. Farther from the seam the iron appeared in both cases less injuriously affected. But although the alternate heating and cooling of the plates over a long period had produced this change in the molecular condition of the iron, a method of restoration presents itself in the process of annealing. In subjecting the pieces cut from the seam-rips to a dull red heat, and then allowing them to cool slowly in sawdust, the writer found that the fibrous character of the iron appeared again, and renewed testing showed that the ductility and tensile strength were restored.

"The same process of annealing is equally effectual in restoring the tenacity of iron in chains rendered brittle, and apparently crystalline, by long use, and is periodically applied where safety depends upon material in this form. Thus the heating and cooling of iron may be looked upon as the bane or the antidote according to the conditions under which the process is carried out. This affords an example of the importance of the physical effects produced by repeated changes of temperature. The change effected by one heating and cooling is so small that a cumulative method of experiment is the only one by which an observable result can be obtained, and this is the method adopted by the writer in the investigation now to be described.

"It is well known that if a wrought-iron bar be heated to redness, a certain expansion takes place, which is most distinctly observed in the direction of its length. It is also known, although not generally so, that if a bar be thus heated and then suddenly cooled in water, a contraction in length takes place, the amount of this contraction exceeding that of the previous expansion, insomuch that the bar when cooled is permanently shorter than it originally was. If this process of heating and cooling be repeated, a further amount of contraction is found to follow for many successive operations.

"Experiments Nos. 1 and 2 were made to verify this, and to show the increment of contraction after each operation.

"EXPERIMENTS ON WROUGHT-IRON BARS 1-1/8 IN. SQUARE BY 30.05 IN. LONG, HEATED TO A DULL RED, THEN COOLED SUDDENLY IN WATER.

------------------+------------------------+------------------------
| EXPERIMENT NO. 1. | EXPERIMENT NO. 2.
| Common Iron. | Best Iron.
+------------+-----------+------------+-----------
| |Percentage | |Percentage
|Contraction.|on original|Contraction.|on original
| | length. | | length.
------------------+------------+-----------+------------+-----------
| Inches. | | Inches. |
After 1st cooling | .04 | .13 | .04 | .13
" 2nd " | .10 | .33 | .10 | .33
" 3rd " | .16 | .53 | .14 | .46
" 4th " | .17 | .56 | .16 | .53
" 5th " | .23 | .76 | .20 | .66
" 6th " | .28 | .93 | .24 | .80
" 7th " | .31 | 1.03 | .27 | .89
" 8th " | .33 | 1.10 | .30 | 1.00
" 9th " | .40 | 1.33 | .33 | 1.10
" 10th " | .47 | 1.56 | .39 | 1.30
" 11th " | .52 | 1.73 | .42 | 1.40
" 12th " | .54 | 1.80 | .47 | 1.56
" 13th " | .58 | 1.93 | .51 | 1.70
" 14th " | .62 | 2.06 | .54 | 1.80
" 15th " | .68 | 2.26 | .56 | 1.86
------------------+------------+-----------+------------+-----------

"The Table of Experiment No. 5 shows that at the twenty-fifth cooling a contraction of 3.05 per cent. had taken place, or an average of .122 per cent. after each cooling. This is almost identically the same average result as shown in Experiment No. 1 with straight bars.

"The above experiments only having reference to the permanent contraction of the iron in the direction of its length, the author made the following experiments to ascertain the effect in the other dimensions, and to see whether the specific gravity of the iron was affected in the reduction of dimensions.

"_Experiment No. 6._--Wrought-iron plate, .74 inch thick, planed on both surfaces and all edges to a form nearly rectangular, and of the dimensions given in Fig. 1429.

"_Specific Gravity._--Two small samples were cut out of different parts of the same piece of plate from which the experimental piece was planed, and the specific gravity determined as follows:--

No. 1 piece 7.629 }
No. 2 piece 7.651 } Mean, 7.64.

"_Quality._--Subjecting a piece to tensile strain in the direction of the grain, it broke at 21.2 tons per square inch of section, the ductility being such that an elongation of 8.3 per cent. occurred before fracture, with a reduction of 9.6 per cent. of the area of fracture. This may be looked upon as representing a fairly good quality of iron.

"A bar of wrought iron, 1-1/8 inches square and 30.00 inches long, was heated to redness, and then allowed to cool gradually in air. Measurements after each of five coolings showed no perceptible change of length.

"_Experiment No. 4._--Wrought-iron bar, 1-1/8 inches square by 30 inches long, heated to a white heat and cooling gradually in air.

------------------+------------+----------------+----------------
|Contraction.| Percentage on | Remarks.
| |original length.|
------------------+------------+----------------+----------------
| Inches. | |
After 1st cooling | No change. | | ----
" 2nd " | " | | ----
" 3rd " | .02 | .07 | ----
" 4th " | .05 | .17 | ----
" 5th " | .05 | .17 | ----
------------------+------------+----------------+----------------

"It may be remarked, that if the bars be heated to white heat a slight contraction does occur, as shown by Experiment No. 4, where a bar of the same dimensions as No. 3 contracted .17 per cent. after the fifth cooling. As, however, the further remarks on this subject have only reference to bars heated to redness and then cooled, the writer would summarize the results of Experiments Nos. 1, 2, and 3, by stating that wrought-iron bars heated to redness permanently contract in their length along the fibre when cooled in water of ordinary temperature; but when cooled in air, they remain unchanged in length.

"To show that this is true as applied to circular hoops, Experiment No. 5 was made upon a wrought-iron bar of 1-1/8 inches square in section, welded into a circular hoop, 57.7 inches outside circumference.

"_Experiment No. 5._--Wrought-iron hoop, 1-1/8 inches square by 57.7 inches outside circumference, heated to a dull red, then cooled suddenly in water.

------------------+------------+--------------+-----------------------
| |Percentage of |
|Contraction.| original | Remarks.
| |circumference.|
------------------+------------+--------------+-----------------------
| Inches. | |
After 1st cooling | .06 | .10 | Red heat.
" 2nd " | .06 | .10 | This was nearly white,
" 3rd " | .16 | .28 | but before cooling
" 4th " | .26 | .45 | red hot.
" 5th " | .35 | .61 |
" 6th " | .46 | .80 |
" 7th " | .54 | .93 |
" 8th " | .60 | 1.04 |
" 9th " | .68 | 1.18 |
" 10th " | .76 | 1.32 |
" 11th " | .80 | 1.38 |
" 12th " | .87 | 1.51 |
" 13th " | .94 | 1.63 |
" 14th " | 1.00 | 1.73 |
" 15th " | 1.08 | 1.90 |
" 20th " | 1.30 | 2.25 | On opposite edge 1.66;
" 25th " | 1.76 | 3.05 | hoop splitting.
------------------+------------+--------------+-----------------------

"This hoop was heated to redness and cooled in water twenty-five times, the circumference of the hoop being accurately measured after each cooling.[23]

[23] The lengths of circumference were taken, in this and other hoops,
after each cooling, by encircling the periphery with a very fine piece
of "crinoline" steel, the ends of which were made just to meet round
the original hoop. By again encircling the hoop with the same piece of
steel the expansion was shown by a gap between the ends, and a
contraction by an overlap, either of which was measured with great
accuracy by means of a finely divided scale.

"Two wrought-iron bars, 1-1/8 inches square and 30.05 inches long, were selected.[24] No. 1 was of common "Crown" quality; No. 2 of a superior quality known as "Tudhoe Crown." These bars were heated to redness in a furnace and then plunged into water of ordinary temperature, the length being accurately measured after each cooling. After fifteen heatings and coolings the permanent contraction on No. 1 bar was 2.26 per cent. of the original length, and that on No. 2 bar 1.86 per cent., or an average on the two bars of about .13 per cent. after each cooling, the increment of contraction being nearly equal after each successive operation. It is noticeable that after the first two coolings the better quality of iron did not contract quite so much as the common quality, and that in the latter the contraction was going on as vigorously at the fifteenth as at the first cooling.

[24] In some of these experiments the original sizes of the iron were
only measured with an ordinary foot-rule, in which case the dimensions
are given in the ordinary fraction used in expressing the mercantile
sizes of iron. When accurate measurement was taken decimals are
invariably used both in this paper and the Tables of Experiment.

"Similar bars of wrought iron, heated to redness and then allowed to cool in air at ordinary temperature, do not appear to suffer any permanent change in their length.

"Experiment No. 3 was made to verify this.

"_Experiment No. 3._--Wrought-iron bar, 1-1/8 inches square by 30 inches long heated to a dull red and cooled gradually in air.

------------------+------------+----------------+-----------
|Contraction.| Percentage on |
| |original length.| Remarks.
------------------+------------+----------------+-----------
After 1st cooling | No change. | ---- | ----
" 2nd " | " | ---- | ----
" 3rd " | " | ---- | ----
" 4th " | " | ---- | ----
" 5th " | " | ---- | ----
------------------+------------+----------------+-----------

(Two-ninths of full size.)

Fig. 1430.]

The plate was subjected to fifty heatings to redness and subsequent coolings in water of ordinary temperature. At every tenth cooling accurate measurements were taken of the contraction in superficial dimensions, and Fig. 1430 shows the final form after fifty coolings. The intermediate measurements at every tenth cooling showed a uniform and gradual decrease in the superficial dimensions, but the thicknesses were only measured after the fifty coolings had been completed. The thickness appears to have varied considerably; in some places, notably towards the centre and outside edges, being much reduced. Between the centre and outside edges the thickness appears to have increased, and in some few places the plate has been split open. The average dimensions in inches before and after the experiment were as follows (dimensions of cracks being allowed for):--

----------------------+---------+----------+----------+-----------
| | | | Cubic
| Average | Average | Average | inches
| length. | breadth. |thickness.|capacity.
+---------+----------+----------+----------
| Inches. | Inches. | Inches. |
Original | 11.995 | 5.98 | .74 | 53.08
After 50 coolings | 11.25 | 5.59 | .774 | 48.72
Per cent. variation { |Decrease |Decrease | Increase | Decrease
from original { | of | of | of | of
{ |6.2 p. c.|6.52 p. c.| 4.6 p. c.| 8.2 p. c.
----------------------+---------+----------+----------+----------

"Three triangular pieces of iron were then cut out of the plate from positions indicated on the diagram; No. 1A from the part most reduced in thickness, No. 3A from the part most increased in thickness, and No. 2A from a part where the thickness was a mean between the thickest and thinnest part. The specific gravities were accurately determined as follows:--

No. 1A 7.552 thinnest part.
No. 2A 7.574 average thickness.
No. 3A 7.560 thickest part.

"The average of these specific gravities is 7.562.

"The average before experiment was 7.64. Hence the average loss in specific gravity has been 1.02 per cent.

"The small triangular piece No. 1A, specific gravity 7.552 (already subjected to fifty heatings when forming part of the solid plate), was next heated and cooled fifty times more. The specific gravity at the end of the one hundred total coolings was 7.52, being .43 per cent. lower than after fifty heatings in plate, and 1.57 per cent. lower than 7.64, the original mean specific gravity of the plate.

"The same piece, 1A, was then heated twenty-five times more, making 125 in all. On taking the specific gravity it was found to be 7.526, or practically the same as after 100 total heatings and coolings.

"It thus appears that there is an undoubted decrease in specific gravity on repeated heating and cooling as described up to one hundred coolings, the specific gravity decreasing as much as 1.57 per cent.; that this percentage appears to be less when the pieces of iron operated upon are very small; that while there is a decrease of specific gravity there is also a decrease of total volume.

"From the above it was evident that the volume was affected by several causes:--

"1. By the permanent contraction of the outer skin, either the volume would be lessened, or relief by bulging out the sides must occur.

"2. By the decrease of specific gravity an increase of volume must occur, which could also find relief in bulging.

"3. A diminution of the whole mass must occur through scaling of the surface.

"Having determined the change in specific gravity by Experiment 6, we only now want to determine the loss of volume due to surface scaling, and we can then infer the actual contraction of the outer skin.

"_Experiment No. 7._--To ascertain the amount of scaling which took place in heating and cooling under same conditions as Experiment No. 6, a wrought-iron plate was cut from the same piece as No. 6, thickness .74 in., planed on both surfaces and all edges to a form nearly rectangular, and to the dimensions given in Fig. 1431.

"The only difference (except the very small difference in the dimensions) between this and 1430, was that the principal grain of the iron was in 1431 in the direction of the arrow, whereas in the other it was lengthwise of the plate.

"This piece was subjected to fifty heatings to redness and sudden coolings in water of ordinary temperature, as in the case of No. 6. The change in form was exactly the same in general character, but the contraction was not quite so great either in length or breadth; the increase in thickness, however, was proportionately greater, the volume (measured by displacement of water) after fifty heatings being 48.6 cubic inches, which is nearly the same as in No. 6 after the same number of heatings. The weight of the piece:--

Avoirdupois.
lbs. oz. dr.
Before heating 14 10 15
After fifty heatings 13 5 10
------------
Difference 1 1 5

"This represents a loss of 9.07 per cent. of the original weight by scaling, and upon the whole original surface (sides and edges) represents a thickness of .0284 of an inch for the fifty immersions, or .00057 of an inch for the thickness of the film lost at each immersion over the whole surface.

"Calculating the weight of No. 6 before and after experiment from the volumes and specific gravities, we find the following:--

Mean Weight of
specific cubic inch
Volume. gravity. water. Pounds.
Weight before heating should be 53.08 × 7.64 × .036 = 14.599
" after " " 48.72 × 7.562 × .036 = 13.262
------
Difference in weight 1.337

the ascertained difference in the case of No. 7 being 1.332, thus sufficiently accounting for the discrepancy between specific gravity and change of volume by the scaling.

"By Experiment 7 it has been shown that the loss of thickness due to scaling after fifty immersions was .0284 inch over the whole surface (sides and edges.) Therefore, assuming this scaling as uniform over the surface, the girth, whether measured lengthwise or breadthwise, should be eight times .0284, or .23 inch less after immersion than before. Now the gross loss of girth is:--

--------------------------------------+-----------+------------
|Lengthwise.|Breadthwise.
--------------------------------------+-----------+------------
| Inches. | Inches.
In No. 6 | 1.38 | .86
In No. 7 | 1.2 | .52
+-----------+------------
Or for both experiments a mean of | 1.29 | .69
Deducting from them the loss of | |
girth due to scaling | .23 | .23
+-----------+------------
Net contraction after fifty immersions| 1.06 | .46
Or in percentage of original girths, | |
which were | 25.46 | 13.43
| per cent. | per cent.
We have a percentage of | 4.16 | 3.42
Or for each immersion an average of | .083 | .07
--------------------------------------+-----------+------------

"Comparing these results with those of Experiments Nos. 1, 2, and 5, we find that the contraction of the skin of the plate is less for each immersion than that of a bar or hoop, in the proportion of .125 to .083. This is what might be expected, as the contraction of the plate is resisted by the volume of heated matter inside, which is eventually displaced by bulging, while the bar finds relief endwise without having to displace the interior.

"We have now before us the following facts, substantiated by the experiments described:--

"1. That in heating to redness, and then cooling suddenly in water at ordinary temperatures, bars and plates of wrought iron, a reduction of specific gravity takes place, the amount being about 1 per cent. after fifty immersions, and 1.57 per cent. after one hundred immersions, further heatings and coolings not appearing to produce further change.

"2. That a reduction of the surface takes place after each heating and cooling, this being due to two causes:--

"_a._ The scaling of the surface, which is shown to amount to a film over the (sides and edges) entire area of .00057 inch in thickness for each immersion, or 0.284 inch for fifty immersions (Experiment 7).

"_b._ A persistent contraction, which takes place after each immersion. This varies according to the form of the iron, being in plates from .07 per cent. to 0.83 per cent. (Experiment 6), while in long bars it varies from .122 to .15 per cent. (Experiments 1, 2, and 5). This contraction continues vigorously up to fifty immersions, and probably much farther.

"3. That in the case of plates a bulging takes place on the largest surfaces, increasing the thickness towards the centres, although the edges diminish in thickness.

"4. That wrought-iron bars heated to redness, and allowed to cool slowly in air, do not show any change in dimensions (Experiment 3).

"The reduction of specific gravity, and the bulging out of the sides, have been explained as follows by the learned Secretary of the Royal Society, Professor Stokes, who has taken considerable interest in these experiments, and who has kindly allowed the author to publish the explanation:

"'When the heated iron is plunged into water, the skin tends everywhere to contract. It cannot, however, do so to any significant extent by a contraction which would leave it similar to itself, because that would imply a squeezing in of the interior metal, which is still expanded by heat, and is almost incompressible. The endeavor, then, of the skin to contract is best satisfied, consistently with the retention of volume of the interior, by a contraction of the skin in the two longish lateral directions, combined with a bulging out in the short direction. The still plastic state of the interior permits of this change.

"'Conceive an india-rubber skin of the form of the plate in its first state, the skin being free from tension, and having its interior filled with water, treacle, or pitch. I make abstraction of gravity. It would retain its shape. But suppose, now, the india-rubber to be endowed with a tension the same everywhere similar to that of india-rubber that has been pulled out, what would take place? Why, the flat faces of considerable area, being comparatively weak to resist the interior pressure, would be bulged out, and the vessel would contract considerably in the long directions, increasing in thickness. This is just what takes place with the iron in the first instance. But when the cooling has made further progress, and the solidified skin has become comparatively thick and strong, the further cooling of the interior tends to make it contract. But this it cannot well do, being encased in a strong hide, and accordingly the interior tends to be left in a porous condition.'

"The reduction by scaling does not require any explanation. The only fact which appears unaccounted for is this persistent contraction of the cooled iron skin, which does not appear to be explicable on any mechanical grounds; and we are, therefore, obliged to look upon it as the result of a change in the distance of the molecules of the iron, caused by the sudden change of temperature in the successive coolings.

"Our next subject is the curious effect of cooling bars or rings by partial immersion in water. Bearing in mind the results at which we have arrived, viz., that wrought iron contracts when immersed in water after heating, and that when allowed to cool in air it remains of the same dimensions, let us ask what would be the behavior of a bar or circular hoop of iron cooled half in water and half in air, the surface of the water being parallel to the fibre and at right angles to the axis of the hoop?

"Arguing from the results of Experiments 1, 2, and 5, it might be expected that the lower portion cooled in water would suffer permanent contraction; and, arguing from Experiment 3, that the upper or air-cooled edge would not alter. This apparently legitimate conclusion is completely disproved by experiments. This will be seen by a reference to Experiments 8, 9, and 10.

"In No. 8 a circular hoop of wrought iron was forged out of a 3-1/2-inch by 1/2-inch bar, the external diameter being about 18 inches, the breadth, 1/2 inch, being parallel to the axis of the hoop. This hoop, Fig. 1432, was heated to redness, then plunged into cold water half its depth, the upper half cooling in air. The changes in the external circumference of the hoop were accurately measured after each of twenty successive coolings, at the end of which the external circumference of the water-cooled edge had increased 1.24 inches, or 2.14 per cent. of its original length, and the air-cooled edge had contracted 7.9 inches, or 13.65 per cent.

"_Experiment No. 8._--Wrought-iron hoop, 3-1/2 inches by 1/2 inch by about 18 inches in diameter, or exactly 57.85 inches in circumference at top, and 57.95 inches at bottom edge.

-------------+--------------------+--------------------+--------------
| Top Edge. | Bottom Edge. |
+--------+-----------+--------+-----------+
|Contrac-|Percentage | Expan- |Percentage |
| tion. |of original| sion. |of original|
| | circum- | | circum- | Remarks.
| | ference. | | ference. |
-------------+--------+-----------+--------+-----------+--------------
| Ins. | | Ins. | |
After 1st dip| .50 | .86 | .08 | .14 |
" 2nd " | .99 | 1.71 | .08 | .14 |
" 3rd " | 1.47 | 2.54 | .26 | .45 |
" 4th " | 1.92 | 3.32 | .30 | .52 |
" 5th " | 2.30 | 3.97 | .34 | .59 |
" 6th " | 2.60 | 4.49 | .40 | .70 |Slight crack
| | | | |in expanded
| | | | |edge.
" 7th " | 2.94 | 5.25 | .44 | .76 |
" 8th " | 3.40 | 5.98 | .50 | .86 |
" 9th " | 3.70 | 6.39 | .56 | .96 |
" 10th " | 4.40 | 7.60 | .62 | 1.07 |
" 11th " | 4.42 | 7.64 | .66 | 1.14 |
" 12th " | 4.85 | 8.40 | .70 | 1.22 |
" 13th " | 5.24 | 9.02 | .78 | 1.34 |
" 14th " | 5.74 | 9.92 | .80 | 1.39 |
" 15th " | 6.00 | 10.37 | .86 | 1.49 |
" 20th " | 7.90 | 13.65 | 1.24 | 2.14 |After de-
| | | | |ducting for a
| | | | |crack .06 inch
| | | | |wide which
| | | | |appeared at
| | | | |sixth dip.
-------------+--------+-----------+--------+-----------+--------------

"It will be observed that we have here two remarkable phenomena: 1. The reversal of the expansion and contraction as described. 2. The very large amount of contraction on the upper edge compared with what was exhibited in Experiment 5 of entire submersion.

"The table showing Experiment 5 gives a contraction of 2.25 per cent. after the twentieth cooling, whereas the contraction on the air-cooled edge of Experiment 8 is 13.65 per cent., or six times the contraction of an entirely submerged hoop.

"To ascertain whether these unexpected phenomena had any connection with the circular form of the hoop, Experiment 9 was made with a straight bar of iron 3-1/2 inches deep by 1/2 inch thick by 28.4 inches long.

"_Experiment No. 9._--Wrought-iron bar, 3-1/2 inches by 1/2 inch by 28.4 inches long, heated to a dull red, then quenched half its depth in water.

------------------+--------------------+--------------------
| Bottom Edge. | Top Edge.
+--------+-----------+--------+-----------
| Expan- |Percentage |Contrac-|Percentage
| sion. |on original| tion. |on original
| | length. | | length.
------------------+--------+-----------+--------+-----------
| Inches.| | Inches.|
After 1st cooling | .05 | .18 | .26 | .91
" 2nd " | .10 | .35 | .43 | 1.51
" 3rd " | .10 | .35 | .54 | 1.90
" 4th " | .14 | .49 | .75 | 2.64
" 5th " | .20 | .70 | .92 | 3.24
" 6th " | .30 | 1.05 | 1.25 | 4.40
" 7th " | .34 | 1.20 | 1.50 | 5.28
" 8th " | .38 | 1.34 | 1.56 | 5.53
" 9th " | .39 | 1.37 | 1.66 | 5.84
" 10th " | .40 | 1.40 | 1.76 | 6.19
" 11th " | .41 | 1.43 | 1.84 | 6.48
" 12th " | .44 | 1.55 | 1.96 | 6.90
------------------+--------+-----------+--------+-----------

"This was cooled half in air and half in water, and the length of the two edges measured accurately after each of twelve coolings. At the end of this experiment the air-cooled edge had contracted 6.9 per cent., while the water-cooled edge had expanded 1.55 per cent. of the original length. The effect on the bar was to make it gradually curve, the water-cooled or extended edge becoming convex, the air-cooled or contracted edge concave.

"Experiment No. 10 was made in order to show the effect of reversing this cooling process. After five coolings, a bar of iron, 28 inches long, 3-1/2 inches deep, and 1/2 inch thick, was curved so that the versed sine of its air-cooled edge was 1-1/2 inches. The coolings were then reversed, what was the air-cooled edge being then immersed in water. After five more coolings the bar was restored to within 1/8 inch of being straight, and the eleventh cooling threw the concavity on the other side of the bar.

"_Experiment No. 10._--Wrought-iron flat bar, 28 inches long by 3-1/2 inches by 1/2 inch, heated to dull red, then quenched half its depth in water, up to five heats, then the opposite edge dipped.

-----------+-----------+------------+------------------
| | | Reversed Cooling.
| | +------------------
|Versed sine| | Versed sine
|of concave,| | of concave,
| _i.e._ | | _i.e._ now
|air-cooled | | water-cooled
| edge. | | edge.
-----------+-----------+------------+------------------
| Inches. | | Inches.
1st cooling| 5/16 | 6th cooling| 1-3/16
2nd " | 9/16 | 7th " | 7/8
3rd " | 13/16 | 8th " | 3/4 scant.
4th " | 1-3/8 | 9th " | 3/8 full.
5th " | 1-1/2 |10th " | 1/8
| |11th | Brought concavity
| | | 1/8 in. on other
| | | side.
-----------+-----------+------------+------------------

"When the author had proceeded thus far, these curious results were shown to several leading scientific men, who expressed interest in the subject, which encouraged the author to extend his experiments under varied conditions with a view of ascertaining the cause for these anomalous effects. These experiments (Nos. 11 to 17) are fully recorded, and the results shown on the diagrams; the actual rings are also on the table before you.

"_Experiment No. 11._--Wrought-iron hoop, turned and bored, 37.1 inches, outside circumference, by 2.95 inches deep by .44 inch thick, the grain of the iron running the short way of the bar from which the hoop was made, heated to redness, then cooled half its depth in water (see Fig. 1437 at A for final form of hoop after ten heatings and coolings).

-----------------+--------------------+--------------------
| Top Edge. | Bottom Edge.
+--------+-----------+--------+-----------
|Contrac-|Percentage | Expan- |Percentage
| tion. |on original| sion. |on original
| | length. | | length.
-----------------+--------+-----------+--------+-----------
| Inches.| | Inches.|
After 1st cooling| .3 | .83 | .05 | .13
" 2nd " | .64 | 1.72 | .12 | .32
" 3rd " | 1.02 | 2.75 | .22 | .60
" 4th " | 1.38 | 3.72 | .30 | .80
" 5th " | 1.62 | 4.37 | .37 | 1.00
" 10th " | 3.14 | 8.46 | .76 | 2.05
-----------------+--------+-----------+--------+-----------

"_Experiment No. 12._--Wrought-iron hoop, turned and bored, 6 inches diameter (18.85 inches circumference) outside, by 2 inches deep by .375 inch thick, heated to redness, then cooled, with lower edge barely touching the water (see Fig. 1437 at B for final form of hoop after twenty heatings and coolings).

-----------------+--------------------+--------------------
| Top Edge. | Bottom Edge.
+--------+-----------+--------+-----------
|Contrac-|Percentage |Contrac-|Percentage
| tion. |of original| tion. |of original
|Outside | circum- |Outside | circum-
|circum- | ference. |circum- | ference.
|ference.| |ference.|
-----------------+--------+-----------+--------+-----------
| Inches.| | Inches.|
After 5th cooling| .10 | .53 | .16 | .85
" 10th " | .22 | 1.17 | .34 | 1.80
" 15th " | .32 | 1.70 | .48 | 2.54
" 20th " | .48 | 2.54 | .62 | 3.30
-----------------+--------+-----------+--------+-----------

"_Experiment No. 13._--Wrought-iron hoop, turned and bored, 6 inches diameter (18.85 inches circumference) outside by 2 inches deep by .375 inch thick, heated to redness, then cooled one-fourth its depth in water (see Fig. 1437 at C for final form of hoop after twenty heatings and coolings).

-----------------+--------------------+--------------------------------
| Top Edge. | Bottom Edge.
+--------+-----------+-------------------+-----------
|Contrac-|Percentage | |Percentage
| tion. |of original| |of original
| | circum- | Extension. | circum-
| | ference. | | ference.
-----------------+--------+-----------+-------------------+-----------
| Inches.| | Inches. |
After 1st cooling| .06 | .32 | .02 | .10
" 5th " | .28 | 1.50 {|A hair's breadth |
| | {|contraction. |
" 10th " | .56 | 3.00 { |Returned to origi- |
| | { |nal circumference. |
" 15th " | .78 | 4.14 | .02 contraction. | .10
" 20th " | 1.12 | 6.00 | .02 contraction. | .10
-----------------+--------+-----------+-------------------+-----------

"_Experiment No. 14._--Wrought-iron hoop, turned and bored. 6 inches diameter (18.85 inches circumference) outside by 2 inches deep by .375 inch thick, heated to redness, then cooled one-half its depth in water (see Fig. 1437 at D for final form of hoop after twenty heatings and coolings).

-----------------+--------------------+--------------------
| Top Edge. | Bottom Edge.
+--------+-----------+--------+-----------
|Contrac-|Percentage | Expan- |Percentage
| tion. |of original| sion. |of original
|Outside | circum- |Outside | circum-
|circum- | ference. |circum- | ference.
|ference.| |ference.|
-----------------+--------+-----------+--------+-----------
| Inches.| | Inches.|
After 5th cooling| .46 | 2.44 | .06 | .32
" 10th " | .96 | 5.00 | .09 | .48
" 15th " | 1.34 | 7.10 | .18 | .96
" 20th " | 1.80 | 9.10 | .26 | 1.38
-----------------+--------+-----------+--------+-----------

"_Experiment No. 15._--Wrought-iron hoop turned and bored, 6 inches in diameter (18.85 inches circumference) outside by 2 inches deep by .375 inch thick, heated to redness, then cooled three-fourths its depth in water (see Fig. 1437 at E for final form of hoop after twenty heatings and coolings).

-----------------+--------------------+-------------------------------
| Top Edge. | Bottom Edge.
+--------+-----------+-------------------+-----------
|Contrac-|Percentage | |Percentage
| tion. |of original| |of original
| | circum- | Expansion. | circum-
| | ference. | | ference.
-----------------+--------+-----------+-------------------+-----------
| Inches.| | Inches. |
After 1st cooling| .05 | .26 | .015 | .08
" 5th " | .30 | 1.60 | .02 | .10
" 10th " | .56 | 3.00 {| A hair's breadth |
| | {| contraction. |
" 15th " | .74 | 3.92 { | .02 |} .10
| | { | contraction. |}
" 20th " | 1.02 | 5.40 {| .03 | } .10
| | {| contraction. | }
-----------------+--------+-----------+-------------------+-----------

"_Experiment No. 16._--Cast-copper ring, turned and bored to same dimensions as Nos. 12, 13, 14, and 15, heated to redness, then cooled half its depth in water (see Fig. 1437 at F for final form of hoop after twenty heatings and coolings).

-----------------+--------------------+--------------------
| Top Edge. | Bottom Edge.
+--------+-----------+--------+-----------
|Contrac-|Percentage | Expan- |Percentage
| tion. |of original| sion. |of original
| | circum- | | circum-
| | ference. | | ference.
-----------------+--------+-----------+--------+-----------
| Inches.| | Inches.|
After 1st cooling| .01 | .05 | .05 | .26
" 2nd " | .01 | .05 | .08 | .42
" 3rd " | .02 | .10 | .14 | .75
" 4th " | .02 | .10 | .17 | .90
" 5th " |} No change from | .22 | 1.17
" 10th " |} original size | .40 | 2.13
" 15th " |} from 5th to | .56 | 3.00
" 20th " |} 20th cooling. | .70 | 3.70
-----------------+--------------------+--------+-----------

"It will be unnecessary to occupy much time in analyzing the experiments, as any one who takes a practical interest in the subject will have full information in the diagrams and tables. Professor Stokes drew attention to the fact that, in 1863, similar phenomena had been noticed by Colonel Clark, of the Royal Engineers. His experiments, made at the Royal Arsenal, Woolwich, were published in the 'Proceedings of the Royal Society,' and Professor Stokes had himself attached an explanatory note, the outline of which was as follows:--

"Imagine a cylinder divided into two parts by a horizontal plane at the water-line, and in this state immersed after heating. The under part, being in contact with water, would rapidly cool and contract, while the upper part would cool but slowly. Consequently by the time the under part had pretty well cooled, the upper part would be left jutting out; but when both parts had cooled their diameters would again agree. Now in the actual experiments the independent motion of the two parts is impossible on account of the continuity of the metal; the under part tends to pull in the upper, and the upper to pull out the under. In this contest the cooler metal, being the stronger, prevails, and so the upper part gets pulled in a little above the water-line while still hot. But it has still to contract in cooling, and this it will do to the full extent due to its temperature, except in so far as it may be prevented by its connection with the rest. Hence, on the whole, the effect of this cause is to leave a permanent contraction a little above the water-line, and it is easy to see that the contraction must be so much nearer to the water-line as the thickness of the metal is less, the other dimensions of the hollow cylinder and the nature of the metal being given. When the hollow cylinder is very short, so as to be reduced to a mere hoop, the same cause operates, but there is not room for more than a general inclination of the surface, leaving the hoop bevelled.

"The expansion of the bottom edge was not noticed in Colonel Clark's paper, perhaps owing to the much smaller hoops which he used in experimenting. Accepting Professor Stokes' explanation of the top contraction, it appears that expansion of the bottom may be accounted for by the reacting strain put on the cooled edge when forcing in the top edge, acting in such a way as to prevent the cooled edge coming quite to its natural contraction, and this, when sufficiently great, expresses itself in the form of a slight expansion.

"_Experiment No. 14._--Forged steel hoop, turned and bored, 18.53 inches in circumference outside by 2.375 inches deep by .27 inch thick, heated to redness, then cooled one-half its depth in water (see Fig. 1437 at G for final form of hoop after three heatings and coolings).

-----------------+--------------------+-------------------+-------------
| Top Edge. | Bottom Edge. |
+--------+-----------+-------+-----------+
|Contrac-|Percentage | Expan-|Percentage |
| tion. |of original| sion. |of original|
| | length. | | length. |
-----------------+--------+-----------+-------+-----------+-------------
| Inches.| |Inches.| |
| | | | {|Cracked at
| | | | {|water-
After 1st cooling| .06 | .32 | -- | -- {|cooled
| | | | {|edge one-
| | | | {|third depth
| | | | {|of ring.
| | | | |
" 2nd " | .12 | .64 | -- | -- |
| | | | |
| | | | {|After allow-
| | | | {|ing for
" 3rd " | .20 | 1.08 | .05 | .27 {|three small
| | | | {|cracks in
| | | | {|bottom edge."
-----------------+--------+-----------+-------+-----------+-------------

The shrinkage of iron and steel by cooling rapidly is sometimes taken advantage of by workmen to refit work, the principles involved in the process being as follows:--

Suppose in Fig. 1438 _a_ _a_ represents a piece of wrought-iron tube that has been heated to a bright red and immersed in cold water _c_ _c_ from the end B to D, _until that end is cold_. The part submerged and cold will be contracted to its normal diameter and have regained its normal strength, while the part above the water, remaining red-hot, will be expanded and weak. There will be, then, a narrow section of the tube, joining the heated and expanded part to the cooled and contracted part, and its form will be conical, as shown at D D. Now, suppose the tube to be slowly lowered in the water, the cold metal below will compress the heated metal immediately above the water-line, the cone section D being carried up into the metal before it has had time to cool; and the tube removed from the water when cold will be as shown in Fig. 1438, from _c_ to D, representing the part first immersed and cooled. To complete the operation the tube must be heated again from the end _c_ to a short distance past D, and then immersed from E nearly to D, and held still until the submerged part is cold, when the tube must be slowly lowered to compress the end _c_ D, making the tube parallel, but smaller in diameter and in bore, while leaving it of its original length, but thickening its wall.

This process may, in many cases, be artificially assisted. Suppose, for example, a washer is too large in its bore; it should have its hole and part of its radial faces filled with fire-clay, as shown in Fig. 1439, in which A is the washer and B B the clay, _c_ _c_ being pieces of wire to hold the fire-clay and prevent its falling off. The washer should be heated to a clear red and plunged in the water D D, which will cool and shrink the exterior and exposed metal in advance of the interior, which will compress to accommodate the contraction of the outer metal, hence the hole will be reduced. This operation may be repeated until the hole be entirely closed.

Another method of closing such a piece as an eye of large diameter compared to its section, is shown in Fig. 1440; first dipping the heated eye at A and holding it there till cold and then slowly lowering it into the water, which would close the diameter across C, and, after reheating, dipping at D till cold, and then slowly immersing, which would close the eye across E. To shrink a square ring, the whole ring would require to be heated and a side of the square dipped, as shown in Fig. 1441, until quite cold, and then immersed slowly for about an inch, the operation being performed with a separate heating for each side. Connecting rod straps, wheel-tires, and a large variety of work may be refitted by this process, but in each case the outside diameter will be reduced.