Statistics For Management And

Economics 9th Edition Keller Solutions

Manual
Visit to download the full and correct content document: https://testbankdeal.com/dow
nload/statistics-for-management-and-economics-9th-edition-keller-solutions-manual/
 Chapter 6
6.1 a Relative frequency approach
b If the conditions today repeat themselves an infinite number of days rain will fall on 10% of the
next days.

6.2 a Subjective approach

b If all the teams in major league baseball have exactly the same players the New York Yankees
will win 25% of all World Series.

6.3 a {a is correct, b is correct, c is correct, d is correct, e is correct}

b P(a is correct) = P(b is correct) = P(c is correct) = P(d is correct) = P(e is correct) = .2
c Classical approach
d In the long run all answers are equally likely to be correct.

6.4 a Subjective approach

b The Dow Jones Industrial Index will increase on 60% of the days if economic conditions remain
unchanged.

6.5 a P(even number) = P(2) + P(4) + P(6) = 1/6 + 1/6 + 1/6 = 3/6 = 1/2
b P(number less than or equal to 4) = P(1) + P(2) + P(3) + P(4) = 1/6 + 1/6 + 1/6 +1/6 = 4/6 = 2/3
c P(number greater than or equal to 5) = P(5) + P(6) = 1/6 + 1/6 = 2/6 = 1/3

6.6 {Adams wins. Brown wins, Collins wins, Dalton wins}

6.7a P(Adams loses) = P(Brown wins) + P(Collins wins) + P(Dalton wins) = .09 + .27 + .22 = .58
b P(either Brown or Dalton wins) = P(Brown wins) + P(Dalton wins) = .09 + .22 = .31
c P(either Adams, Brown, or Collins wins) = P(Adams wins) + P(Brown wins) + P(Collins wins)
= .42 + .09 + .27 = .78

6.8 a {0, 1, 2, 3, 4, 5}
b {4, 5}
c P(5) = .10
d P(2, 3, or 4) = P(2) + P(3) + P(4) = .26 + .21 + .18 = .65
e P(6) = 0

6.9 {Contractor 1 wins, Contractor 2 wins, Contractor 3 wins}

157
6.10 P(Contractor 1 wins) = 2/6, P(Contractor 2 wins) = 3/6, P(Contractor 3 wins) = 1/6

6.11 a {Shopper pays cash, shopper pays by credit card, shopper pays by debit card}
b P(Shopper pays cash) = .30, P(Shopper pays by credit card) = .60, P(Shopper pays by debit card)
= .10
c Relative frequency approach

6.12 a P(shopper does not use credit card) = P(shopper pays cash) + P(shopper pays by debit card)
= .30 + .10 = .40
b P(shopper pays cash or uses a credit card) = P(shopper pays cash) + P(shopper pays by credit
card)
= .30 + .60 = .90

6.13 {single, divorced, widowed}

6.14 a P(single) = .15, P(married) = .50, P(divorced) = .25, P(widowed) = .10

b Relative frequency approach

6.15 a P(single) = .15

b P(adult is not divorced) = P(single) + P(married) + P(widowed) = .15+ .50 + .10 = .75
c P(adult is either widowed or divorced) = P(divorced) + P(widowed) = .25 + .10 = .35

6.16 P( A1 ) = .1 + .2 = .3, P( A 2 ) = .3 + .1 = .4, P( A 3 ) = .2 + .1 = .3.

P( B1 ) = .1 + .3 + .2 = .6, P( B 2 ) = .2 + .1 + .1 = .4.

6.17 P( A1 ) = .4 + .2 = .6, P( A 2 ) = .3 + .1 = .4. P( B1 ) = .4 + .3 = .7, P( B 2 ) = .2 + .1 = .3.

P(A1 and B1 ) .4
6.18 a P(A1 | B1 ) = = = .57
P(B1 ) .7

P(A 2 and B1 ) .3
b P(A 2 | B1 ) = = = .43
P(B1 ) .7

c Yes. It is not a coincidence. Given B1 the events A1 and A 2 constitute the entire sample space.

P(A1 and B 2 ) .2
6.19 a P(A1 | B 2 ) = = = .67
P( B 2 ) .3

158
 P(A1 and B2 ) .2
b P( B 2 | A 1 ) = = = .33
P(A1) .6
c One of the conditional probabilities would be greater than 1, which is not possible.

6.20 The events are not independent because P(A 1 | B 2 )  P(A 1 ) .

6.21 a P( A 1 or B1 ) = P(A 1 ) + P(B1 ) − P(A 1 and B1 ) = .6 + .7 - .4 = .9

b P( A 1 or B 2 ) = P(A 1 ) + P(B 2 ) − P(A 1 and B 2 ) = .6 + .3 - .2 = .7

c P( A 1 or A 2 ) = P(A 1 ) + P(A 2 ) = .6 + .4 = 1

P(A1 and B1 ) .20

6.22 P(A1 | B1 ) = = = .25 ; P(A 1 ) = .20 + .05 = .25 ; the events are
P(B1 ) .20 + .60

independent.

P(A1 and B1 ) .20

6.23 P(A1 | B1 ) = = = .571 ; P(A 1 ) = .20 + .60 = .80 ; the events are
P(B1 ) .20 + .15

dependent.

6.24 P( A1 ) = .15 + .25 = .40, P( A 2 ) = .20 + .25 = .45, P( A 3 ) = .10 + .05 = .15.

P( B1 ) = .15 + 20 + .10 = .45, P( B 2 ) = .25 + .25 + .05 = .55.

P(A 2 and B 2 ) .25

6.25 a P( A 2 | B 2 ) = = = .455
P( B 2 ) .55

P(A 2 and B 2 ) .25

b P( B 2 | A 2 ) = = = .556
P( A 2 ) .45

P(A 2 and B1 ) .20

c P( B1 | A 2 ) = = = .444
P( A 2 ) .45

6.26 a P( A 1 or A 2 ) = P( A 1 ) + P( A 2 ) = .40 + .45 = .85

b P( A 2 or B 2 ) = P( A 2 ) + P( B 2 ) – P( A 2 and B 2 ) = .45 + .55 - .25 = .75

c P( A 3 or B1 ) =P( A 3 ) + P( B1 ) – P( A 3 and B1 ) = .15 + .45 - .10 = .50

P(promoted and female ) .03

6.27 a P(promoted | female) = = = .20
P(female ) .03 + .12

P(promoted and male ) .17

b P(promoted | male) = = = .20
P(male ) .17 + .68

159
c No, because promotion and gender are independent events.

6.28 a P(debit card) = .04 + .18 + .14 = .36

P(credit card and over $100 .23
b P(over $100 | credit card) = = = .49
P(credit card) .03 + .21 + .23
c P(credit card or debit card) = P(credit card) + P(debit card) = .47 + .36 = .83

6.29 a P(Less than high school) = .077 + .110 = .187

b P(college/university | female) =
P(college / university and female ) .092
= = .198
P(female ) .077 + .154 + .141 + .092

P(high school and male ) .201

c b P(high school | male) = = = .566
P(male ) .154 + .201

6.30 a P(He is a smoker) = .12 + .19 = .31

b P(He does not have lung disease) = .19 + .66 = .85
c P(He has lung disease | he is a smoker) =
P(he has lung disease and he is a smo ker) .12
= = .387
P(he is a smo ker) .31
d P(He has lung disease | he does not smoke) =
P(he has lung disease and he does not smoke) .03
= = .043
P(he does not smoke) .69

6.31 The events are dependent because P(he has lung disease) = .15, P(he has lung disease | he is a
smoker) = .387

P(manual and math − stats) .23

6.32 a P(manual | math-stats) = = = .390
P(math − stats) .23 + .36
b P(computer) = .36 + .30 = .66
c No, because P(manual) = .23 + .11 = .34, which is not equal to P(manual | math-stats).

6.33 a P(customer will return and good rating) =.35

P(good rating and will return) .35 .35
b P(good rating | will return) = = = = .538
P( will return) .02 + .08 + .35 + .20 .65

P(good rating and will return) .35 .35

c P(will return| good rating) = = = .714
P(good rating ) .35 + .14 .49

d (a) is the joint probability and (b) and (c) are conditional probabilities

160
6.34 a P(ulcer) = .01 + .03 + .03 + .04 = .11
P(ulcer and none) .01 .01
b P(ulcer | none) = = = = .043
P(none) .01 + .22 .23

P(ulcer and none) .01 .01

c P(none | ulcer) = = = = .091
P(ulcer ) .01 + .03 + .03 + .04 .11

P(ulcer and none)

d P(One, two, or more than two | no ulcer) = 1 − = 1 − .091 = .909
P(ulcer )

P(Insufficient work and 25 − 54) .180

6.35 a P(Insufficient work | 25-54) = = = .252
P(25 − 54) .320 + .180 + .214

b P(65 and over) = .029 + .011 + .016 = .056

c P(65 and over |plant or company closed or moved) =
P(65 and over and plant or company closed or moved) .029
= = .064
P(plant or company closed or moved) .015 + .320 + ..089 + .029

6.36 a P(remember) = .15 + .18 = .33

P(remember and violent ) .15 .15
b P(remember | violent) = = = = .30
P( violent ) .15 + .35 .50
c Yes, the events are dependent.

P(above average and murderer) .27 .27

6.37 a P(above average | murderer) = = = = .563
P(murderer) .27 + .21 .48
b No, because P(above average) = .27 + .24 = .51, which is not equal to P(above average
testosterone | murderer).

6.38a P(Health insurance) = .167+.209+.225+.177 = .778

b. P(Person 55-64 | No health insurance)
P(Person 55 − 64 and No health insurance) .026 .026
= = = = .128
P(Person 55 − 64) .177 + .026 .203

c. P(Person 25-34|No health insurance) =

P(Person 25 − 34 and No health insurance) .085 .085
= = = .385
P( No health insurance) .085 + .061 + .049 + .026 .221

P(Violent crime and primary school) .393 .393

6.39a = = = .673
P(Pr imary school) .393 + .191 .584

b. P(No violent crime) = .191 + .010 + .007 + .015 = .223

161
 P(Violent crime and enrollment less than 300) .159 .159
6.40a = = = .636
P(Enrollment less than 300) .159 + .091 .250

P(Violent crime and enrollment less than 300) .159 .159

b. = = = .205
P(Violent crime) .159 + .221 + .289 + .108 .777

P(new and overdue) .06 .06

6.41 a P(new | overdue) = = = = .103
P(overdue) .06 + .52 .58

P(new and overdue) .06 .06

b P(overdue | new) = = = = .316
P(new ) .06 + .13 .19
c Yes, because P(new) = .19  P(new | overdue)

6.42 a P(under 20) = .464 + .147 + .237 = .848

b P(retail) = .237 + .035 + .005 = .277
P(20 to 99 and construction ) .039 .039
c P(20 to 99 | construction) = = = = .077
P(construction ) .464 + .039 + .005 .508

6.43 a P(fully repaid) = .19 + .64 = .83

P(fully repaid and under 400) .19 .19
b P(fully repaid | under 400) = = = = .594
P(under 400) .19 + .13 .32

P(fully repaid and 400 or more) .64 .64

c P(fully repaid | 400 or more) = = = = .941
P(400 or more) .64 + .04 .68
d No, because P(fully repaid)  P(fully repaid | under 400)

P(purchase and see ad ) .18 .18

6.44 P(purchase | see ad) = = = = .30; P(purchase) = .18 +
P(see ad ) .18 + .42 .60
.12 = .30. The events are independent and thus, the ads are not effective.

6.45 a P(unemployed | high school graduate) =

P(unemployed and high school graduate) .014 .014
= = = .047
P(high school graduate) .282 + .014 .296

b P(employed) = .091 + .282 + .166 + .095 + .213 + .115 = .962

c P(advanced degree | unemployed) =
P(advanced deg ree and unemployed ) .002 .002
= = = .053
P(unemployed ) .008 + .014 + .007 + .003 + .004 + .002 .038

d P(not a high school graduate) = .091 + .008 = .099

6.46 a P(bachelor’s degree | west)

162
 P(bachelor ' s deg ree and west) .046 .046
= = = = .201
P( west) .036 + .059 + .045 + .020 + .046 + .023 .229
b P(northeast | high school graduate)
P(northeast and high school graduate ) .063 .063
= = = = .199
P(high school graduate ) .063 + .078 + .117 + .059 .317
c P(south) = .059 + .117 + .061 + .030 + .065 + .032 = .364
d P(not south) = 1 –P(south) = 1−.364 = .636

6.47

6.48

163
6.49

6.50

6.51

164
6.52

a P(R and R) = .81

b P(L and L) = .01
c P(R and L) + P(L and R) = .09 + .09 = .18
d P(Rand L) + P(L and R) + P(R and R) = .09 + .09 + .81 = .99

6.53 a & b

c 0 right-handers 1
1 right-hander 3
2 right-handers 3
3 right-handers 1
d P(0 right-handers) = .001

165
P(1 right-hander) = 3(.009) = .027
P(2 right-handers) = 3(.081) = .243
P(3 right-handers) = .729

6.54a

b P(RR) = .8091
c P(LL) = .0091
d P(RL) + P(LR) = .0909 + .0909 = .1818
e P(RL) + P(LR) + P(RR) = .0909 + .0909 + .8091 = .9909

6.55a

P(0 right-handers) = (10/100)(9/99)(8/98) = .0007

166
P(1 right-hander) = 3(90/100)(10/99)(9/98) = .0249
P)2 right-handers) = 3(90/100)(89/99)(10/98) = .2478
P(3 right-handers) = (90/100)(89/99)(88/98) = .7265

6.56

a P(win both) = .28

b P(lose both) = .30
c P(win only one) = .12 + .30 = .42

6.57

P(sale) = .04

167
6.58

P(D) = .02 + .018 = .038

6.59

P(Same party affiliation) = P(DD) + P(RR) + P(OO) = .1936 + .1369 + .0361.3666

168
6.60

Diversity index = .12 + .04 + .12 + .0075 + .04 + .0075 = .335

6.61

P(heart attack) = .0504 + .0792 = .1296

169
6.62

P(pass) = .228 + .243 + .227 = .698

6.63

P(good ) = .3132 + .0416 = .3548

170
6.64

P(myopic) = .1008 + .1512 = .2520

6.65

P(does not have to be discarded) = .1848 + .78 = .9648

6.66 Let A = mutual fund outperforms the market in the first year
B = mutual outperforms the market in the second year
P(A and B) = P(A)P(B | A) = (.15)(.22) = .033

6.67 Let A = DJIA increase and B = NASDAQ increase

P(A) = .60 and P(B | A) = .77
P(A and B) = P(A)P(B | A) = (.60)(.77) = .462

6.68 Define the events:

M: The main control will fail.
B1: The first backup will fail.
B2: The second backup will fail

171
The probability that the plane will crash is
P(M and B1 and B2) = [P(M)][ P(B1)][ P(B2)]
= (.0001) (.01) (.01)
= .00000001
We have assumed that the 3 systems will fail independently of one another.

6.69 P( wireless Web user uses it primarily for e-mail) = .69

P(3 wireless Web users use it primarily for e-mail) = (.69)(.69)(.69) = .3285

6.70

P(Increase) = .05 + .5625 = .6125

6.71 P(A and B) = .36, P(B) = .36 + .07 = .43

P(A and B) .36
P(A | B) = = = .837
P(B) .43

6.72 P(A and B) = .32, P(AC and B) = .14, P(B) = .46, P(BC) = .54
P(A and B) .32
a P(A | B) = = = .696
P(B) .46

P(A C and B) .14

b P(AC | B) = = = .304
P(B) .46

P(A and B C ) .48

c P(A and BC) = .48; P(A | BC ) = C
= = .889
P( B ) .54

P(A C and B C ) .06

d P(AC and BC) = .06; P(AC | BC) = C
= = .111
P( B ) .54

172
6.73

P(B) = .4940 + .0115 = .5055

P(A and B) .4940
P(A | B) = = = .9773
P(B) .5055

P(F and D) .020

6.74 P(F | D) = = = .526
P(D) .038

6.75 Define events: A = crash with fatality, B = BAC is greater than .09)
P(A) = .01, P(B | A) = .084, P(B) = .12
P(A and B) = (.01)(.084) = .00084
P(A and B) .00084
P(A | B) = = = .007
P(B) .12

P(CFA I and passed) .228

6.76 P(CFA I | passed) = = = .327
P(passed) .698

6.77 Define events: A = heart attack, B = periodontal disease

P(A) = .10, P(B | A) = .85, P(B | AC ) = .29

173
P(B ) = .085 + .261 = .346
P(A and B) .085
P(A | B) = = = .246
P(B) .346

6.78 P(A) = .40, P(B | A) = .85, P(B | AC ) = .29

P(B ) = .34 + .174 = .514

P(A and B) .34
P(A | B) = = = .661
P(B) .514

6.79 Define events: A = smoke, B1 = did not finish high school, B 2 = high school graduate, B 3 =

some college, no degree, B 4 = completed a degree

P(A | B1 ) = .40, P(A | B 2 ) = .34, P(A | B 3 ) = .24, P(A | B 4 ) = .14

From Exercise 6.45: P( B1 ) = .1055, P( B 2 ) = .3236, P( B 3 ) = .1847, P( B 4 ) = .3862

174
P(A) = .0422 + .1100 + .0443 + .0541 = .2506
P( B 4 | A) = .0541/.2506 = .2159

6.80 Define events: A, B, C = airlines A, B, and C, D = on time

P(A) = .50, P(B) = .30, P(C) = .20, P(D | A) = .80, P(D | B) = .65, P(D | C) = .40

P(D) = .40 + .195 + .08 = .675

P(A and D) .40
P(A | D) = = = .593
P(D) .675

6.81 Define events: A = win series, B = win first game

175
P(A) = .60, P(B | A) = .70, P(B | AC ) =.25

P(BC ) = .18 + .30 = .48

P(A and B C ) .18

P(A | BC ) = C
= = .375
P(B ) .48

6.82

P(PT) = .28 + .052 = .332

P(R and PT) .28
P(R | PT) = = = .843
P(PT) .332

176
6.83

P(PT) = .0046 + .0269 = .0315

P(H and PT) .0046
P(H | PT) = = = .1460
P(PT) .0315

6.84 Sensitivity = P(PT | H) = .920

Specificity = P(NT | H C ) = .973

Positive predictive value = P(H | PT) = .1460

P(H C and NT) .9681 .9681

Negative predictive value = P H C | NT) = = = = .9996
P( NT) .0004 + .9681 .9685

6.85

177
P(PT) = .0164 + .6233 = .6397
P(NT) = .0036 + .3567 = .3603
P(C and PT) .0164
P(C | PT) = = = .0256
P(PT) .6397

P(C and NT) .0036

P(C | NT) = = = .0010
P( NT) .3603

6.86 a P(Marketing A) = .053 + .237 = .290

b P(Marketing A | Statistics not A) =
P(Marketing A and Statistics not A) .23 .237
= = = .290
P(Statistics not A) .237 + .580 .817
c Yes, the probabilities in Parts a and b are the same.

6.87 Define events: A = win contract A and B = win contract B

a P(A and B) = .12

b P(A and BC) + P(AC and B ) = .18 + .14 = .32
c P(A and B) + P(A and BC ) + P(AC and B ) = .12 + .18 + .14 = .44

6.88 a P(second) = .05 + .14 = .19

P(successfuland − 8 or less ) .15 .15
b P(successful | –8 or less) = = = = .517
P(−8 or less ) .15 + .14 .29

c No, because P(successful) = .66 + .15 = .81, which is not equal to P(successful | –8 or less) .

178
6.89 Define events: A = woman, B = drug is effective

P(B) = .528 + .221 = .749

P(A C and B) .221

6.90 P(AC | B) = = = .295
P(B) .749

6.91 P(Idle roughly)

= P(at least one spark plug malfunctions) = 1– P(all function) = 1 – (.90 4 ) = 1-.6561 = .3439

6.92

P(no sale) = .65 + .175 = .825

6.93 a P(pass) = .86 + .03 = .89

P(pass and miss 5 or more classes) .03 .03
b P(pass | miss 5 or more classes) = = = = .250
P(miss 5 or more classes) .09 + .03 .12

P(pass and miss less than 5 classes) .86 .86

c P(pass | miss less than 5 classes) = = = = .977 .
P(miss less than 5 classes) .86 + .02 .88

d No since P(pass)  P(pass | miss 5 or more classes)

179
6.94 Define events: R = reoffend, D = detained

a P(D) = P(R and D) + P(R C and D) = .1107 + .2263 = .3370

P(R and D) .1107
P(R| D) = = = .3285
P(D) .3370

b P(D C ) = P(R and D C ) + P(R C and D C ) = .1593 + .5037 = .6630

P(R and D C ) .1593

P(R| D C ) = C
= = .2403
P( D ) .6630

6.95 a P(excellent) = .27 + .22 = .49

P(man and excellent ) .22 .22
b P(excellent | man) = = = = .44
P(man ) .22 + .10 + .12 + .06 .50

P(man and excellent ) .22 .22

c P(man | excellent) = = = = .449
P(excellent ) .27 + .22 .49

d No, since P(excellent)  P(excellent | man)

180
6.96

P(R) = .0176 + .5888 = .6064

P(S and R ) .5888
P(S | R) = = = .9710
P( R ) .6064

6.97 Define events: A1 = Low-income earner, A 2 = medium-income earner, A 3 = high-income

earner, B = die of a heart attack, BC survive a heart attack

P(BC ) = .1848 + .4459 + .2790 = .9097

P(A 1 and B C ) .1848

P( A1 | BC ) = C
= = .2031
P( B ) .9097

181
6.98 Define the events: A1 = envelope containing two Maui brochures is selected, A 2 = envelope

containing two Oahu brochures is selected, A 3 = envelope containing one Maui and one Oahu

brochures is selected. B = a Maui brochure is removed from the selected envelope.

P(B) = 1/3 + 0 + 1/6 = 1/2

P(A1 and B) 1 / 3
P( A1 | B) = = = 2/3
P(B) 1/ 2

6.99 Define events: A = purchase extended warranty, B = regular price

P(A and B) .21 .21
a P(A | B) = = = = .2692
P(B) .21 + .57 .78
b P(A) = .21 + .14 = .35
c No, because P(A)  P(A | B)

6.100 Define events: A = company fail, B = predict bankruptcy

P(B) = .068 + .2392 = .3072

182
 P(A and B) .068
P(A | B) = = = .2214
P(B) .3072

6.101 Define events: A = job security is an important issue, B = pension benefits is an important
issue
P(A) = .74, P(B) = .65, P(A | B) = .60
a P(A and B) = P(B)P(A | B) = (.65)(.60) = .39
b P(A or B) = .74 + .65 – .39 = 1

6.102 Probabilities of outcomes: P(HH) = .25, P(HT) = .25, P(TH) = .25, P(TT) = .25
P(TT | HH is not possible) = .25/(.25 + .25 + .25) = .333

6.103 P(T) = .5

Case 6.1
1. P(Curtain A) = 1/3, P(Curtain B) = 1/3
2. P(Curtain A) = 1/3, P(Curtain B) = 2/3
Switch to Curtain B and double your probability of winning the car.

Case 6.2
Probability Bases Probability Joint
Outcome of outcome Occupied Outs of scoring Probability
1 .75 2nd 1 .42 .3150
2 .10 1st 1 .26 .0260
3 .10 none 2 .07 .0070
4 .05 1st and 2nd 0 .59 .0295
P(scoring) = .3775
Because the probability of scoring with a runner on first base with no outs (.39) is greater than the
probability of scoring after bunting (.3775) you should not bunt.

Case 6.3
0 outs:
Probability of scoring any runs from first base = .39
Probability of scoring from second base = probability of successful steal  probability of scoring
any runs from second base = (.68)(.57) = .3876
Decision: Do not attempt to steal.

183
1 out:
Probability of scoring any runs from first base = .26
Probability of scoring from second base = probability of successful steal  probability of scoring
any runs from second base = (.68)  (.42) = .2856
Decision: Attempt to steal.

2 outs:
Probability of scoring any runs from first base = .13
Probability of scoring from second base = probability of successful steal  probability of scoring
any runs from second base = (.68)  (.24) = .1632
Decision: Attempt to steal.

Case 6.4

Age 25: P(D) = 1/1,300

P(D and PT) = (1/1,300)(.624) = .00048
P(D and NT) = (1/1,300)(.376) = .00029

P( D C and PT) = (1,299/1,300)(.04) = .03997

P( D C and NT) = (1,299/1,300)(.96) = .95926
P(PT) = .00048 + .03997 = .04045
P(NT) = .00029 + .95926 = .95955
P(D | PT) = .00048/.04045 = .01187
P(D | NT) = .00029/.95955 = .00030

184
Age 30: P(D) = 1/900
P(D and PT) = (1/900)(.710) = .00079
P(D and NT) = (1/900)(.290) = .00032

P( D C and PT) = (899/900)(.082) = .08190

P( D C and NT) = (899/900)(.918) = .91698

P(PT) = .00079 + .08190 = .08269
P(NT) = .00032 + .91698 = .91730
P(D | PT) = .00079/.08269 = .00955
P(D | NT) = .00032/.91730 = .00035

Age 35: P(D) = 1/350

P(D and PT) = (1/350)(.731) = .00209
P(D and NT) = (1/350)(.269) = .00077
P( D C and PT) = (349/350)(.178) = .17749

P( D C and NT) = (349/350)(.822) = .81965

P(PT) = .00209 + .17749 = .17958
P(NT) = .00077 + .81965 = .82042
P(D | PT) = .00209/.17958 = .01163
P(D | NT) = .00077/.82042 = .00094

Age 40: P(D) = 1/100

P(D and PT) = (1/100)(.971) = .00971
P(D and NT) = (1/100)(.029) = .00029

P( D C and PT) = (99/100)(.343) = .33957

P( D C and NT) = (99/100)(.657) = .65043
P(PT) = .00971 + .33957 = .34928
P(NT) = .00029 + .65043 = .65072
P(D | PT) = .00971/.34928 = .02780
P(D | NT) = .00029/.65072 = .00045

Age 45: P(D) = 1/25

P(D and PT) = (1/25)(.971) = .03884
P(D and NT) = (1/25)(.029) = .00116

P( D C and PT) = (24/25)(.343) = .32928

P( D C and NT) = (24/25)(.657) = .63072
P(PT) = .03884 + .32928 = .36812

185
P(NT) = .00116 + .63072 = .63188
P(D | PT) = .03884/.36812 = .10551
P(D | NT) = .00116/.63188 = .00184

Age 49: P(D) = 1/12

P(D and PT) = (1/12)(.971) = .08092
P(D and NT) = (1/12)(.029) = .00242

P( D C and PT) = (11/12)(.343) = .31442

P( D C and NT) = (11/12)(.657) = .60255

P(PT) = .08092 + .31442 = .39533
P(NT) = .00242 + .60255 = .60467
P(D | PT) = .08092/.39533 = .20468
P(D | NT) = .00242/.60467 = .00400

Case 6.5
The probability that 23 people have different birthdays is .4927. The probability that at least two
people have the same birthday is 1 − .4927 = .5073.

186
Another random document with
no related content on Scribd:
 As an invariable rule the rate of cooling fixes the degree of hardness to be had
in the cold piece within the limits of obtainable hardness or softness.
Slow cooling retains softness, so that when annealing is to be done the slower
the cooling the better. Cooling is always a hardening process, but when it is
carried on slowly more softness, will be retained than when the cooling is quick.
Rapid cooling produces hardness, and the more nearly instantaneous it is the
greater the hardness will be. This property of hardening is of such extreme
importance that it will be treated fully in a separate chapter.
There is an apparent exception to this rule shown in the operation called
water-annealing. It is common, when work is hurried, to heat a piece of steel
carefully and uniformly up to the first color, that is, until it just begins to show
color, and then to quench it in water.
This is called water-annealing; and many believe that because a piece so
treated is left softer than it was before treatment, the water-cooling had
something to do with it. The fact is that hammering and rolling are hardening
processes. When the increment of heat due to the work is less than the
decrement of heat due to radiation, the compacting of the grain increases
hardness.
This process leaves the piece harder than does the quenching in water-
annealing; the decrease in hardness due to water-annealing is the difference
between the effects of the two operations. Let two pieces of the same bar be
heated exactly the same for water-annealing; let one be quenched in water, and
the other be allowed to cool in the air in a dry place. Then the superior softness of
the air-cooled piece will show that the so-called water-annealing furnishes no
exception to the rule.
There is one extremely important matter connected with cooling that should be
noted carefully.
It is a common practice among steel-workers when they get a part of a piece
of steel too hot to partially quench that part, and then go on with their heating; or
if they are in a hurry to get out a big day’s work, or if the weather is hot, and a pile
of red-hot bars is uncomfortable, to dash water over the pile and hurry the
cooling.
This practice means checks in the steel, hundreds of them.
 A bar breaks and has this appearance. The dark spot is the check; it did not
show in the bar, no inspector could see it, but it broke the bar. Any one can prove
this to his own satisfaction in a few minutes. Take a bar of convenient size, about
one inch by one eighth; heat it carefully to an even medium orange color and
quench it completely; then snip it with a hand-hammer over the edge of an anvil,
snipping away until satisfied that it is sound steel. There are no checks.
Now beat a similar length of the same bar in the same way, and pass it
through the stream from the bosh-pipe, or submerge it for a moment in the bosh,
not long enough to produce more than the slightest trace of a change in the color;
then put it back in the fire and bring it gently to the uniform color used before, and
quench it completely. Now when it is snipped over the anvil it will show numerous
checks, dozens of them.
In this experiment the complete submersion for a moment may not produce
checks at every trial, because the complete submersion permits practically
uniform cooling, which if continued to complete cooling would be simply the
ordinary hardening process. Still it will produce checks in the majority of cases,
indicating that starting the changes, strains, or whatever they are of the
quenching process and then stopping them suddenly while the steel is in the
plastic condition does cause disintegration, so that the operation is dangerous
and should not be tolerated. Passing the hot steel through a stream of water or
dashing water over it must cause different rates of cooling, and necessarily
produce local strains resulting in checks. These latter ways of injuring, therefore,
rarely fail to produce the ruinous checks.
If this positive destruction is produced in this way, in steel containing enough
carbon to harden it is clear that similar, although not so pronounced, results will
be produced in the mildest steels when they are treated in the same manner.
 The rule, then, should be: Never allow water to come in contact with hot steel,
and never allow hot steel to be laid down upon a damp floor.
Even the spray from water which is run upon roll-necks may cause these
checks in steel that is passing through the rolls, so that it is better to put up a
guard to deflect such water away from the body of the roll.
A hammerman may sweep a bar with a damp broom to cause the vapor to
explode with violence when the hammer comes down, and so tear away all rough
scale and produce a beautiful finish. A careful, skilful man may be permitted to do
this, but as surely as he gets his broom too wet, so that drops of water will fall on
the steel and whirl around in the spheroidal condition, just so surely will he check
the steel.
The best way is to have the broom not wet enough to drip, and then to strike it
up against the top die when it is ready to descend; sufficient moisture will be
caught upon the die to cause a loud explosion when it strikes the hot steel; it is a
violent explosion and will drive off every particle of detachable scale, leaving as
beautiful a surface as that which is peculiar to Russia sheet iron.
It is common in rolling tires to run jets of water over the tire to break up the
scale and produce a clean surface. Tire-makers assert that experience shows
that the water does no harm. There are two reasons for this if it be true: first, the
steel is of medium carbon and more inert than high steel, and it has been
hammered and compacted before rolling; second, the tires are usually turned,
and this would cut away any little checks that might occur on the surface.
The magnetic properties of steel are well known. Soft steel, like soft wrought
iron, cannot be magnetized permanently; higher carbon steel will retain
magnetism a long time, and hardened steel will retain it still longer. Hardened-
steel magnets are the most permanent.
The permanency and the efficiency of a magnet increase with the quantity of
carbon up to about 85 carbon; steel of higher carbon than this will not make
magnets of so good permanency. The efficiency of a magnet of 85 carbon is
increased largely by the addition of a little tungsten; a little less than .05% is
sufficient.
It has been shown that tungsten has the property of retaining the hardness of
steel up to a relatively high temperature; this additional power of retaining
magnetism may indicate a close relation between the conditions set up by
magnetism and by hardening.
It has been stated that maximum physical properties, except as to
compression, are found at from 90 to 100 carbon; now we find maximum
magnetic properties in the same region. Prof. Arnold has found by microscopic
tests the same point of saturation; he fixes it at 89 carbon and deduces from it an
unstable carbide of Fe₂₄C.
 The magnetic maximum was found by magnet-makers by actual use in large
numbers of magnets. Prof. J. W. Langley found the same maximum in a series of
careful and delicate experiments undertaken to determine the best composition
and the best treatment for the production of permanent magnets. Magnetism is
affected by temperature, and it is found that steel becomes non-magnetic at or
about the point of recalescence. This is important to electricians, as it marks the
limit of temperature that is available to them. It is of interest to the scientists, as it
is another indication of the importance of the changes that take place at this
temperature. Later, recalescence will be found to be an equally important point to
the steel-worker, especially to the temperer.
It has been stated that if a bar of steel be heated to any visible temperature
and then be cooled without disturbance there will be a resulting grain or structure
that is due to the highest temperature to which the bar was subjected. As a rule
the highest temperature leaves a grain that appears to the eye to be the largest,
or coarsest, whether the microscope shows it to be composed of larger crystals or
not.
Let the following squares represent the apparent sizes of the grains:

1. The natural bar, untreated.

2. Grain due to dark orange or orange red.
3. “ “ “ medium orange.
4. “ “ “ bright orange.
5. “ “ “ dark lemon.
6. “ “ “ medium lemon.
7. “ “ “ bright lemon.
8. “ “ “ very bright lemon, or creamy.

These designations are used because steel in cooling down, or in heating up,
runs through a series of yellow tints, not reds. It is common to see the expression
“glowing white” applied to steel that is not even melted, when as a matter of fact
melted wrought iron is not quite white. An occasional heat of steel may be seen
that could fairly be called white, and then the melter knows that it is altogether too
hot, and that he must cool the steel or make bad ingots. “Glowing white,” like
“cherry red,” will do for ordinary talk, but not for accurate description, although
“cherry red” comes nearer to describing the dying color than “glowing white”
comes to describing the highest heat.
An arc light may be “glowing white,” and sunlight is “glowing white,” and when
either light falls upon melted steel it shows how far the steel is from being
“glowing white.”
Referring to the squares: If a bar that has been heated to No. 8 be re-heated
to No. 2 and be kept at that color a few minutes to allow the steel to arrange itself,
in other words, to provide for lag, and then be cooled, it will be found to have
grain No. 2. Sometimes in performing this experiment the fracture will be
interspersed with brilliant spots as if it were set with gems; this shows that not
quite enough time was allowed for lag. Another trial with a little more time will
bring it to a complete No. 2 fracture. If now it be heated to No. 4, or 5, or 6 in the
same way, it will be found to have when cold the grain due to No. 4, or 5, or 6
temperature.
This may be repeated any number of times, and the changes may be rung on
all of the numbers, until the disintegrating effect of numerous heatings begins to
destroy the steel. This property of registering temperature, this steel thermometer,
is of great value, and it will be referred to frequently.

EFFECTS OF MECHANICAL WORK.

When an ingot is heated and then hammered, rolled, or pressed hot, its
density will be increased, as well as its strength when cold under all strains.
If it be hammered carefully, with heavy blows at first, and with lighter and
quicker blows at the last, the grain will become very close and fine; it is called
“hammer-refined.”
When down to the so-called cherry red, orange red, great care is needed, and
when black begins to show through the red much caution must be used; any
heavy blows will crush the grain and produce the dark or black color mentioned
before.
Fine-tool makers attach great importance to this hammer-refining; some of the
most expert will not have a rolled bar if a well-hammered one can be had. At first
thought this would seem to be a mere notion, but the testimony in favor of
hammering is so universal among those who know their business that it would
seem as if it must be based upon some reason. If it have any scientific basis of
fact, it is that the shocks or vibrations of the hammer keep the carbon in more
intimate union with the iron, whether it be combination or solution, than either
rolling or pressing will do. After considering the phenomena of hardening,
tempering, annealing, etc., it may be concluded that there is something in this. It
is easy to laugh at and to deride shop prejudices, and there are enough of them
that deserve ridicule; again, there are some that will not down, and they compel
the scientist to hunt for explanations. But after all, ridicule is dangerous; it is
possible that a careful comparison of some of the laws laid down by the highest
scientists would tend to excite the risibles. If the hand-worker sometimes
flounders in the mud, the scientist is sometimes enveloped and groping in mist.
 Hot-rolling produces results similar to those of hot-hammering; it makes the
grain finer, increases density, and adds to the strength.
The same precautions are needed in rolling as in hammering. Heavy passes
with rapid reduction may be used to advantage while the steel is hot and
thoroughly plastic; as the heat falls the passes should be lighter to avoid crushing
the grain.
Overrolling, like too much hammering, may be more injurious than too little
work; a coarse, irregular structure due to too little work may be rectified and made
fine and even by annealing, while if the grain be crushed by overwork the
damage cannot be cured by annealing; the annealed grain may appear to be all
right, but on testing, the strength will be found impaired.
By care and light passes steel may be rolled safely down to a black heat and
be made elastic and springy. It is common to roll spring-steel in this way so that it
may be formed into a spring and have all of the properties of a tempered spring
without going through the operations of hardening and tempering. This is often
desirable for spring-makers, as it saves them considerable expense; but it is
hazardous work, because it is so difficult to heat every piece exactly to the same
temperature, and secure every time the same number of passes and the same
pressure in each. The best roller will get some pieces too hard and brittle, and
some too soft and ductile. A careful steel-maker will shun such work.
Cold-hammering, cold-rolling, and cold-drawing reduce specific-gravity and
increase tensile, transverse, compressive, and torsional strength. They increase
hardness and brittleness, reducing ductility. The hardness due to cold-working is
different from that due to hot-work or quenching; the latter operations produce
great elasticity as well as hardness.
The hardness due to cold-working might be described as harshness; the steel
is not truly springy; of course it will bend farther without permanent set than an
annealed piece, but it never has the true spring elasticity. If it be worked far
enough to be really springy, it will bear the same relation to a hot-worked spring
that a piece of cross-grained, brashy oak bears to a piece of well-seasoned,
straight-grained hickory.
The hammering of round sections between flat dies tends to burst the bars in
the centre; great care must be used to avoid this, and the most skilful and careful
hammermen will often turn out bursted bars. The bursts do not show on the
surface; the bars are true to size, round, smooth, and sound on the outside. The
safest plan is to hammer in a V-die, or in rounded swedges.
Radial rolling will produce the same results, and it is on this principle that the
celebrated Mansmann tubes are made. The explanation seems to be simple, as
the following exaggerated sketches will show:
 No. 1 has been struck; it is then turned up to position No. 2 and knocked into
shape No. 3. The rapid hammering of a bar, turning it a little at a time, must burst
it if the blows are heavy enough to deform the whole section. Heavy radial rolling
produces the same results.
The concluding pages of this chapter will be devoted to a few examples
showing by tests the effects of heat and work upon specific-gravity, tensile
strength, elasticity, and ductility; they are not to be taken as fixing exact limits in
any case; they are given merely to illustrate the truth of the general properties
stated, and to show the wide ranges of strength that are attainable by varying
carbon and work.

TABLE I
Character Ingot Numbers.
of
Steel. 1 2 3 4 5 6 7 8 9 10 11 12
.302 .490 .529 .649 .801 .841 .867 .871 .955
Carbon 1.079
1.005 1.058
Silicon .019 .034 .043 .039 .029 .039 .057 .053 .059 .088 .120 .039
Phosphorus .047 .005 .047 .030 .035 .024 .014 .024 .070 .034 .064 .044
Sulphur .018 .016 .018 .012 .016 .010 .018 .012 .016 .012 .006 .004
Sp. gr. 7.807 7.808
7.805
ingots. 7.855 7.836 7.841 7.829 7.838 7.824 7.819 7.818 7.813
Sp. gr.
bars,
burned, 1 7.818 7.791 7.789 7.752 7.744 7.690
2 7.814 7.811 7.784 7.755 7.749 7.741
3 7.823 7.830 7.780 7.758 7.755 7.769
4 7.826 7.849 7.808 7.773 7.789 7.798
5 7.881 7.806 7.812 7.790 7.812 7.811
cold, 6 7.844 7.824 7.829 7.825 7.826 7.825
Diff. 6-1 .025 .034 .040 .073 .082 .135
Mean diff.
.071
of carbon
 The twelve ingots treated here were first selected by ocular inspection for
carbons; the carbons were then determined by combustion analyses.
It will be seen that the inspection was correct, and that the mean difference in
carbon between consecutive numbers is .007. Between Nos. 7 and 8 there is a
difference of only .004; when the analyst discovered this, he asked for a
reinspection, not giving any reason for his request. The inspectors made new
fractures, examined the ingots carefully in good light, and reported that they erred
the first time, that both ingots belonged in the same temper number, but that if
there were any difference No. 8 was the harder. It is not claimed that a difference
of .004 is really observable.
The contents of silicon, phosphorus, and sulphur show clearly that the
controlling element is carbon. This experiment has been repeated a number of
times, and always with the same result, showing that there is no uncertainty in
this method of separating tempers.
Parts of these ingots were reduced to ¾-inch round bars. The specific
gravities of the ingots were taken, showing generally a reduction of sp. gr. for an
increase of carbon. No. 3 and 5 are anomalous; an explanation of this could
doubtless have been found if a careful investigation had been made, but there
was no re-examination.
The sp. gr. No. 6 are of the ¾-inch bars as they came from the rolls; they are
all heavier than the ingots except No. 4, and they are of nearly uniform sp. gr.;
this is due doubtless to the fact that the higher carbon steels are so much harder
than the low-carbon steels that it required much more work to reduce them to the
bars, and as hot-working increases density, the densities of the higher carbons
were increased more than those of the lower.
The bars were nicked six times at intervals of about ¾-inch and then heated
so that the ends were scintillating, ready to pass into the granular condition, and
the heat was so regulated as to have each piece less hot than the piece next
nearer to the end, the last piece, No. 6, being black and as nearly cold as
possible.
It is manifest that this operation is subject to the error of accidentally getting
No. 2, for instance, hotter than No. 1, and so on, so that perfect regularity is not to
be expected; to obtain a true rule of expansion it would be necessary to make
hundreds of such experiments and use the mean of all.
It will be noticed that No. 4 is abnormal in the ingot series, and that the No. 6
piece of No. 4 is abnormal in being lighter than the ingot; probably this No. 6 of
No. 4 was hot when it was intended to be cold. Also No. 2 of ingot No. 3 is lighter
than its No. 1, showing another irregularity in heating.
Taking the whole list of No. 1 pieces, they are all lighter than their respective
No. 6 pieces; the differences of sp. gr. 6-1 are progressive, being only .025 for the
No. 3 ingot and .135 for the No. 12 ingot. This shows clearly that expansion due
to a given difference in temperature is much greater in high steel than in low
steel.
This clears away the mystery of the so-called treachery of high steel, its
tendency to crack when hardened. There is no treachery about it; it is very
sensitive to temperature, and it must be treated accordingly.
A few examples will now be given to show the changes of tensile strength,
ductility, etc., that may be had by differences of carbon, and by differences of
treatment, annealing, hardening, and tempering.

TABLE II
Crucible
Character Crucible Crucible Crucible
Crucible ½-in.
of O. H. O. H. O. H. O. H. Eye-bar, Eye-bar, Eye-bar,
Sheet Drawn
Steel. 2″ × 1″. 2″ × 1″. 2″ × 1″.
Wire.
.09 to .96 1.15
Carbon .435 .50 .60 .70 1.35 1.40
.12
Silicon .008 .014 .025 .156 < .02
Phosphorus .007 .050 .016 .008 < .02
Sulphur .026 .028 .028 .015 trace
Manganese .055 .204 .325 .24 < .30
Tensile
strength,
lbs. per sq. 46800 73142 84220 124800 100733 117710 141500
108800 117400
in.
Elastic limit 30900 63560 71500 69980 65000 85087 69850 92420
in 2 in 1 25% 14.5% 11.5% 4.75% .5% 7.28 at 2%
Elongation
in. 41% in. 2.85 in
42% 2½
75.85% 62.3% 29.91% 13.03% 2.42%
Reduction of
13.55% 8.59%
area

broke
in neck broke broke
silky
Fracture slight in head in
½
cup flaw, close grip
fine grain
grain

O. H. is the abbreviation for open hearth.

Second column is mean of 24 analyses and 24 tests of boiler-sheets.
 TABLE III
Tensile Elastic Elongation. Reduction
Cold-drawn Wire, Strength, Limit,
of Area,
½-inch Diam. lbs. per lbs. per In 3 Per
in. cent. per ct.
sq. in. sq. in.
Cold-drawn, broke in grip 141,500 92,400 .06 2.00 2.42
Same bar drawn black 138.400 114,700 .18 6.00 12.45
“ “ annealed 98,410 68,110 .30 10.00 11.69
“ “ hardened and then drawn black 248,700 152,800 .25 8.33 19.7

Analysis of this bar is given in Table II in the last column.

A test of ½-inch wire to show effect of cold-drawing, tempering, annealing, and
hardening and tempering. Four pieces were cut from the same bar. It is probable
that the first piece would have given a little higher tensile if it had not broken in
the grip; it was clamped too tight. The second piece was heated until it passed
through all of the temper colors and turned black, technically called “drawing
black,” or drawing out all of the temper. It is not quite annealing; the idea was to
find the effect of temper-drawing upon a cold-hardened drawn wire.
The effect of this operation was to lower the ultimate and raise the elastic
strength, increasing also the ductility.
The third piece was heated carefully to the recalescence-point, and cooled
slowly, thus annealing it completely, and giving the normal strength of a bar of this
composition.
The fourth piece was heated to recalescence and quenched, hardening and
refining it thoroughly; it was then tempered through all of the colors until it turned
black; the result shows the enormous potencies there are in the hardening and
tempering operations.
The cases given in Table II were selected indiscriminately, so as to show
better the effect of carbon, as we here have tests of ordinary test-bars, boiler-
sheet, small eye-bars, and drawn wire.
The 96-carbon eye-bar and the 115-carbon ½-inch wire are the nearest to the
100-carbon saturation limit mentioned before, and they show the highest
strength. The 96-carbon eye-bar had a slight flaw in the fracture, which doubtless
caused it to break below its real strength.
The 135-carbon eye-bar broke in the head in a way to indicate that there was
some local strain there, due to forging.
These examples are not given as establishing any general law; they are
illustrations of what all experience shows to be the fact, that the strength of steel
is affected profoundly by the quantity of carbon present, and also by heat and by
mechanical work. From 46,800 lbs. to 248,700 lbs. tensile strength per square
inch is an enormous range, and these figures probably represent pretty closely
the ultimate limits at present attainable.
An inspection of the analyses makes it clear that the other elements present in
addition to carbon were not there in sufficient quantity or variety to have had
much effect upon the results.
 VI.
HEATING FOR FORGING; FOR HARDENING;
FOR WELDING.

BURNING, OVERHEATING, RESTORING.

From what has been said already about the effects of heat it
follows without further argument that heating is one of the most
important, or perhaps more properly the most important of all, of the
operations to which steel has to be subjected.
The first and vital thing to be borne in mind is that all heating
should be uniform throughout the mass. It has been shown that heat
affects the grain, the structure, as surely as it moves the mercury-
column, and such being the case it is plain that as perfect uniformity
as it is possible to attain is the first essential for all heating, no matter
what the ultimate object may be.
In heating for forging the limit lies between the point of
recalescence, the beginning of true plasticity, and the granular
condition, the end of plasticity; these temperatures lie between dark
or medium orange for all steels and medium or light lemon on the
upper limit, depending on the carbon content, or lower if it be an
alloy steel.
If there is much work to be done upon a piece of steel, it is well to
heat at first to as high a temperature as is safe, and then to forge or
work heavily at the higher heat, reducing the blows or passes as the
piece is reduced and the temperature falls. Although this high
heating will raise the grain of the steel, the heavy working will bring it
back to a fine, compact structure.
If little work is to be done, then it is better to heat as low as may
be safe, and allow the work to be done without letting the heat down
below orange red, so that the steel may not be crushed in the grain.
Below orange red, the so called “dark cherry,” steel should not be
forged, except that in forging for fine tools it is well to give many light
and rapid blows until black begins to show in order to hammer-refine
it; this must be done with extreme care so as not to crush the steel
and cause cracking in the subsequent hardening, or crumbling in the
hardened tool.

HEATING FOR HARDENING.

When a piece of steel is to be hardened by quenching in water or
any quick-cooling medium, it should be heated with great care to the
exact temperature to produce the required hardness.
After forging, no piece of steel should be quenched without first
being heated uniformly to the proper temperature. Ede in his book
recommends quenching immediately after forging in some cases.
The so-called Harvey patent recommends cooling from a high heat
down to the required heat and then quenching.
Both practices are bad. In the Ede case this is believed to be the
only bad piece of advice in his very valuable book—in every other
respect the most practical and useful book upon the manipulation of
steel known to the author.
The reason for objecting to the quenching after forging without re-
heating is that forging always sets up uneven strains in the mass; the
flow is easier from the sides than from the middle of the piece, and
therefore the amount of work done upon one part is greater than
upon another; also it is impossible to hammer or press a piece of
steel with exact uniformity throughout, so that it follows that after
forging there is never exact uniformity of texture or temperature, and
such uniformity is the one essential thing to insure good and even
hardening.
The practice of allowing a highly heated piece to cool down to a
given color and then quenching is objectionable, because it
produces a coarse and brittle grain due to the higher heat.
 Referring to the illustration on page 67 of the squares
representing grains due to different temperatures: Assume that
square No. 3 represents the heat at which quenching is to take
place, and No. 6 is the heat to which the piece has been subjected;
then the piece when it has cooled to No. 3 will not have the grain due
to No. 3 heat: it will have a larger, coarser grain that formed as the
piece cooled from No. 6. If now it be quenched, it will have only the
hardness due to No. 3, with a much coarser and more brittle grain
than No. 3 heat should give. The way to manage such a case is to
let the piece cool completely and assume the No. 6 grain; then re-
heat carefully to exactly No. 3 and no hotter; keep the piece at that
heat for a few minutes, or moments, according to its size, to allow for
lag; then it will have the finer grain due to No. 3 heat, and when
quenched it will be as hard as under the other method, and it will be
much finer and stronger.
The same rule applies to any two temperatures.
As an expression of exactness as to evenness of heat, it may be
said that the piece should be as uniform in color as if it had been
dipped into a pot of paint. When such uniformity is attained, a break
from quenching is rare, unless the piece has been shamefully
overheated so that the strains of quenching are greater than the
tenacity of the steel.

HEATING FOR WELDING.

When an ingot is to be forged or rolled, it is well to take the
highest heat possible—that immediately below the heat of
granulation. Such a heat may be taken safely by keeping the steel
covered with a surface flux to protect it from the flame. Ordinary red
clay, dried and powdered, is an excellent flux for the purpose, and
the cheapest known. Melted and powdered borax is the best of
known fluxes, but it is so expensive that, as a rule, it is used only on
the finest tool-steel, or on some of the alloy steels where the highest
heat possible is not above a bright orange color, or hardly so high.
 A good flux, intermediate in cost between common red clay and
powdered borax, is an earth or mineral barite, or heavy spar. This
material fuses more readily than red clay and not quite so easily as
borax. It forms a good protective covering on the steel, and it is
nearly or quite as efficient as borax.
The object in heating so high is to make the steel as soft and
plastic as it may be, so that the subsequent working will close up all
porosity as far as possible. Nearly all ingots have in them a greater
or less number of cavities, commonly called blow-holes, that are
caused by the separation of occluded gases during cooling. If such
porosities are not oxidized on the surface they will disappear under
heavy working at a high heat. It is probable that under the
compression of the work the gases are redisseminated in the mass
and the walls of the cavities are reunited. If there be the slightest
oxidation of the surface of a cavity the walls will not reunite: there will
be left in the mass a little flat film of oxide which will prevent the
union.
In mild steels used for machinery or structural purposes these
little films may do no harm, the factor of safety being sufficient to
more than cover any weakening effect. In tool-steel that is to be
hardened such little films are almost certain to cause fracture. Dies
as large as twelve inches square and six to eight inches thick, having
been heated and quenched with the greatest care, have split fairly in
two, and have revealed in the fracture a little film no larger than half
an inch in diameter and of inappreciable thickness. At the same time
the perfectly uniform grain and hardness showed that the highest
skill had been used. This is only one illustration of the fact that every
break in the continuity of the grain in steel forms a starting-point for
fracture under heavy stress.
From what has been said it is plain that to weld two pieces of
steel together is a difficult matter; still it can be done if great care be
used. In general it is better to avoid such welding except in cases of
necessity. The welding of steel tubing, and the electric welding of
rails, frogs, switches, etc., is done on a large scale and satisfactorily,
so that it will not do to say that steel cannot be welded. It can be
welded or pasted together, and it is a good operation to avoid in all
high steel. In case steel is to be hardened a weld will reveal itself
almost certainly.

BURNING IN HEATING.
When a piece of steel breaks and shows a coarse, fiery fracture,
it is common to say that it is burned. This is not necessarily the case.
There are several degrees in the effects of heat. The first is the
raising of the grain; the second, in high steel, is the decarbonizing or
burning out of carbon from the surface in, the depth of the
decarbonizing depending upon time and temperature; the third is
oxidizing, or actual burning in the common acceptance of the term.
All of these operations go on to a slight extent every time a piece
of steel is heated, but when the heating is done carefully there is
only a small film of steel that is decarbonized and oxidized, and this
film flies off when the piece is quenched for hardening. When the
steel is forged or rolled this skin will be united firmly to the steel, and
it will be thinner or thicker, according to the number of heatings and
the time of exposure to the fire. In tool-making this skin must always
be removed. Many an expensive tool is made perfectly worthless by
not having this skin all removed, owing usually to mistaken economy.
The steel is expensive, and the tool-maker does not wish to cut it up
into worthless chips.
When a tool costing, say, twenty-five dollars is made useless by
failure to cut away twenty-five cents’ worth of useless skin, the
economy of such an operation requires no discussion. It is
impossible to forge a piece of steel without producing such a skin,
and it is well known that decarbonized iron will not harden.
Ordinarily a cut of ¹/₁₆ of an inch should remove such a skin on
straight rolled or hammered bars. In the case of a shaped forging
where many re-heatings have been required the forgeman will have
done good work if the cutting away of ⅛ of an inch will present a
good surface: tool-makers should consider this and allow for it. On
the other hand, if a tool-maker finds that the removal of ⅛ of an inch
from a bar, or ¼ of an inch from a forging will not yield him a good,
hard surface, he should hold the steel-maker responsible for bad
work.
Actual burning reveals itself in rough tears, and cracks at the
surface and corners of the piece. Such a piece should go to the
scrap heap.
Overheated steel of coarse, fiery grain has been injured, and not
necessarily destroyed. Such a piece may be restored to any fineness
of grain by heating to the right temperature—medium orange for the
best grain—keeping it at that heat for, say, one minute for a little
piece, and five to ten or fifteen minutes for a large piece. The heat
should penetrate the whole mass, and it should not be allowed to run
above the given color in any part, not even for a moment. It should
then be allowed to cool in a dry place, without disturbance. The grain
will now be fine and uniform, and the steel may be worked in the
ordinary way.
This simple operation is all that is necessary to restore to a fine
grain any piece of steel that has been overheated, provided that the
piece has not been actually burned nor ruptured.
 VII.
ANNEALING.

It has been shown that the grain or structure of steel is profoundly

affected by heat, so that any difference of heat-color that is visible to
the naked eye will cause a difference of grain that is also visible to
the naked eye.
Specific-gravity tests and delicate magnetic tests have proved
that for every variation in grain there is a difference of specific-
gravity, which means, of course, a difference in volume; from this it is
clear that if in any one piece of steel there exists a variety of grain
due to uneven heating, there must necessarily be in the mass
internal destructive strains. These strains become manifest when a
piece of unevenly heated steel cracks in hardening; in this case the
strains are greater than the tenacity of the steel.
It is well known, also, that all working of steel, such as forging or
rolling, has a hardening effect, so that ordinary bars or forgings
cannot be machined readily in the condition in which they are left by
these operations.
If there were no remedy for these conditions of internal stress
and initial hardness, the general use of steel would be very difficult,
and its application would be limited seriously.
Fortunately, there are three properties of steel which furnish an
easy and efficient remedy.
First, the fact that steel will assume by mere heating a grain or
structure due to any temperature, no matter what its previous
structure may have been, makes it a simple matter to remove
practically all irregularities of grain and stress, by heating the mass
to a perfectly uniform color and allowing it to cool uniformly.
 Second, as heating is a softening process always, the mere
heating of any piece of steel will soften it, and the amount of this
softness that can be retained when the piece is cold is a direct
function of the length of time of cooling, so that by sufficiently slow
cooling any steel can be left reasonably soft.
This does not apply to Hadfield’s manganese steel, which cannot
be made soft when cold by any of the known processes of
annealing.
Third, by reference to the specific-gravity table No. I, Chap. V, it
will be seen that the change in volume due to differences of
temperature is much less in mild steel than in high steel. This fact
does not rest upon the evidence of this table alone; it is a fact of
common knowledge to all steel-makers that mild steel is much more
inert than high steel; therefore differences of heat and working that
produce serious results in high steel are hardly appreciable in mild
steel. As a rule all structural steels are comparatively mild, therefore
they are generally in a fit condition for use when they leave the rolls
or forge. In cases of special forging, where one part is heated and
another is left cold, as in the forging of the heads of eye-bars, it
would seem to be wiser to anneal such pieces to remove the area of
strain that must exist between the unheated parts and those that
were heated and forged.
The operation of removing strains and hardness by careful,
uniform heating and slow cooling is known as annealing.
Annealing should not be confused with tempering. Tempering is
the partial softening of hardened steel, to remove some of the
exceeding brittleness of hardened steel, and so to make it strong
and highly elastic while it is still very hard.
Annealing is the complete softening of a piece of steel; that is to
say, as a rule, the obtaining of the utmost softness that is possible;
or in any case to have the steel softer than any tempering would
leave it.
Annealing, and tempering are frequently used synonymously.
Such misuse of terms in speaking of technical matters leads to
confusion of ideas and misunderstandings.
As a rule, the best heat to use for annealing is that which gives a
medium orange color; it is a good heat to quench from; it is a little
above the heat of recalescence, about 655° Cent. This heat is that
which gives the finest grain to steel when it is hardened, and is
known as the refining heat.
As steel is thoroughly plastic and soft at this heat, and as it yields
the best and strongest grain when cooled from this heat, it is clear
that there is nothing to be gained by heating any higher for
annealing.
In annealing, the steel should be brought up to the right color,
medium orange, and left at that heat until it is hot through, care being
taken that the heat does not run any higher in any part of the piece.
If the corners or edges or any part be allowed to run up to bright
orange, or to medium or bright lemon, as is often done, then there is
bad work; the result will be uneven grain and internal strains.
When steel is to be hardened afterwards, there may be no harm
in heating up to an even lemon color; but where is the use in
applying this excess of heat merely to make a coarse grain, when
the lower, medium orange color will give just as good softness and a
much better grain?
The time necessary for good annealing depends upon the size of
the piece; a wire may be brought up to the right heat in five minutes
or less, and heated through in another minute; then it should be
removed from the fire, as every additional moment of heating will
only injure the steel.
A block six or eight inches cube may require three to five hours to
bring it up to the color and have it heated through, and sufficient time
should be given; but as soon as it is hot through it should be
removed from the fire.
A six-inch block may be brought up to a medium orange color in
twenty minutes or less in a hot furnace, and then if it be kept in such
a furnace until it is hot all through, the surface and edges will almost
certainly be brought to a bright lemon color, with bad results. To do