Sie sind auf Seite 1von 6

Mathematical Association of America

Fractions without Quotients: Arithmetic of Repeating Decimals


Author(s): Richard Plagge
Source: The Two-Year College Mathematics Journal, Vol. 9, No. 1 (Jan., 1978), pp. 11-15
Published by: Mathematical Association of America
Stable URL: http://www.jstor.org/stable/3026549
Accessed: 27-10-2015 06:22 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

Mathematical Association of America is collaborating with JSTOR to digitize, preserve and extend access to The Two-Year
College Mathematics Journal.

http://www.jstor.org

This content downloaded from 131.91.169.193 on Tue, 27 Oct 2015 06:22:29 UTC
All use subject to JSTOR Terms and Conditions
Mathematical

FractionsWithoutQuotients:Arithmetic
of
RepeatingDecimals
Richard Plagge
Richard Plagge has taughtat Manteno High School, Manteno,
Illinois,and is currentlyan instructorin Mathematics at High-
line CommunityCollege, Midway, Washington, where he has
been since 1967. He earned his B.S. in Education at Illinois
o .s State in 1962 and his M.S. in Mathematics at Western Washing-
t22ba
ton State College in 1967.

It is well known(and repeatedlytaught)thata real numberis rationalif and onlyif


it can be writtenas an infiniterepeatingdecimal.That the decimal representation
is not necessarilyunique is also well known.However,if we do not allow those
representations withrepeatingzeroes(oftencalled terminating), therepresentations
are unique. It is also truethat thereis a one-to-onecorrespondencebetweenthe
non-zero real numbers and all non-terminating decimals (repeating or not).
Strangely, zero is the onlyintegerwhichcannotbe writtenwithrepeatingnines.
While discussingthese ideas and the propertiesof real numbersin an in-
termediatealgebra class, I decided to point out why we would assume the field
propertiesof the real numbers.We had proved that multiplicationand addition
wereclosed operationson the rationalnumbers.Why shouldn'twe at least do this
much for real numbers?Since the only representationsavailable to us of a
computationalnature were the infinitedecimals (the number line is not very
computationaland Dedekind cuts were threemilesover theirheads), I decided to
give theclass some idea of the difficulties
encounteredin definingand performing
additionand multiplication of real numbers.To thisend, I asked if theycould add
or multiplyrational numbersusing only the decimal representations and their
knowledgeof integers.
This can be a shakingexperienceforsome studentswho are veryconfidentthey
have mastered"simplearithmetic"like adding and multiplying. It is well,then,to
help themdiscoverhow to add repeatingdecimals.Also, to keep thingssimple,we
consideredonlypositivenumbers.My class had alreadyseen how to convertfrom
fractionsto decimals and fromrepeatingdecimals to fractions.They were aware
that terminatingdecimals could be writtenwith repeatingnines. We used the
notation of underscoring a block of digits which repeat. For example:
1/3 =3 =.33 =.33333.
We looked at some exampleswhichgave themno trouble:.3 +.4 = 1/3 + 4/9
= 7/9 =.7 Then some a little more difficult:
(a) .6+.5=6/9+5/9=11/9=1.2 (not1.1)
(b) .73 +.58 = 73/99 + 58/99 = 131/99 = 1.32 (not 1.1)
(c) 1.35 + 2.57 = 122/90 + 232/90 = 354/90 = 3.93 (not 3.92)
(d) 1.03 + 2.7 = 1.03 + 2.77 = 93/90 + 250/90 = 343/90 = 3.81 (not 3.80)
The studentssoon saw thatif we must"carry"beyond the block of digitswhich

11

This content downloaded from 131.91.169.193 on Tue, 27 Oct 2015 06:22:29 UTC
All use subject to JSTOR Terms and Conditions
repeats,we mustadd one to theblock whichthenrepeatsin thesum. Note thatwe
are studyinga binaryoperationso we need onlyconcernourselveswithaddingtwo
numbersat a time.Thus, the only numberwe ever carryis one! That should give
you some idea how to extendthisdiscussionto adding morethantwo numbers.
In the previousexamples,the lengthof the blocks of the numberswe were
addingwas always the same. So, considernow: .054 +.98565
=.05454545454 +.978565985659
= 1.04020531114.
Reason: 5454545454+ 8565985659= 14020531113.
ten digits ten digits ten digits

Notice we mustfirstget the repeatingblocks to startat the same place and also
have themthe same length.
An obvious theorem: If A has repeatingblocks n digitsin lengthand B has
repeatingblocksm digitsin lengththenA + B has repeatingblocks L.C.M. (n, m)
digitsin length.
With the previousdiscussionone can easily formulatean algorithmfor adding
repeatingdecimals.Here's how it can be done:
Instructions: Example:
(A) Writetheproblem. (A) 2.70 584
+ 6.91749
(B) Rewritetheproblemso thatbothhave (B) 2.70 584584
repeatingblocksof thesame length. + 6.917 494949
(See "obvious theorem.")
(C) Rewriteproblemagain so thatrepeating (C) 2.705 845845
blocks"start"at same position. + 6.917 494949
(D) Add as you ordinarilywould terminating (D-F) 9.623340794
decimalsadd I to thelast digitifand +1
onlyifyou "carried"past theblock 9.623 340795
whichrepeats.
(E) Underscorethelast digitsso thatthe
numberof underscoreddigitsis the
same as in part(B).
(F) Simplifyifpossible.
Now considersome multiplication problems;
(A) .3 x.4 = x =
3/9 4/9 12/81 =.148
(B) .1 x.2 = 1/9 x 2/9 = 2/81 =.024691358!!!
Thinkyourstudentswill findan algorithmformultiplication? Mine didn't.Let us
findone. Using the associativelaw and the commutativelaw:

(A) ( O(.4) = (3)(4)(1 /9)2 = (12)(1 /9)2 = (12)(10 -1)-2

12

This content downloaded from 131.91.169.193 on Tue, 27 Oct 2015 06:22:29 UTC
All use subject to JSTOR Terms and Conditions
(B) (-1)(.2) = (1)(2)(1 /9)2 = (2)(1 /9)2 = (2)(10- 1)- 2
(C) (.37)(.54) = (37)(54)(1/99)2 = 1998(102 _ 1) - 2
(D) (.102)(.102) =(102/999)(101/990) = (102102/999999)(1020201/9999990)
=(102102)(1020201)(106- I)-2(10)-1
=(104164562502)(106 - )2(10)- ,

or betteryet

(.102)(.102) = (.102)(1 + .02)(10)-' = [.102 + (.102)(.02)](10)-I


and (.102)(.02) = (102/999)(2/99) = (102102/999999)(020202/999999)
= (102102)(20202)(106 - 1)-2 = 2062664604(106 - 1)-2.

Thus to do multiplicationof repeatingdecimals we must be able to multiply


integers,findintegermultiplesof repeatingdecimals,and be able to evaluate the
form (1011-1)2.
Let us see how we can multiplyan integerby a repeatingdecimal.Here are two
examples:
(A) 2358 X .67= (2358)(67)(.01)
= (157986)(.0_)
= 01010.10
+ 0505.05
+ 070.70
+ 09.09
+ 0.80
+ .06
= 01595.80
+.01 (since 1 was carriedpast the block)
1595.81

(B) 90102i X.01= 010101 . 0101 X.01


198 digits
=01010101 . . .010101.01 i= 98
+010101 . . .010101.01 i= 97
+ 0101 ... 010101.01 i= 96
+ 01 ... 010101.01 i= 95

+ 6ioioi.oi i = 3
+ 0101.01 i= 2
+ 01.01 i= 1
+ .01 i=0
01020304 . . . 969798.99= 0102030405... 969799.
There is a reasonfordoing thepreviousrathermessyexample.We need to be able
to evaluate(10' - 1)-2. Let us do thisforn = 2.

13

This content downloaded from 131.91.169.193 on Tue, 27 Oct 2015 06:22:29 UTC
All use subject to JSTOR Terms and Conditions
(102 1)-2= (1/99)2= (. 01)2= (. 01)(1/99) =. 01 X 010101 ... 01
999999 ... 99
98
196 digits

i=O 01020304 . .. 969799


01 98 =.0001020304 ... 969799
9999999999 . . . 9 99 99 9
E98

i =(1 198digits
In a similarmannerone may show the following:
(10 - 1) -2 = ( 1/9)2 = (. 1)2 =.012345679
(102 _ 1)2 = (1/99)2 = (.l )2=. 00 01 02 03 ... 95 96 97 99

(10 _ 1)-2 =(1/999)2 =(.00)2 =.00001 002 ... 996 997 999

(104 - 1)-2 = (1/9999)2= (. 000)2=.0000 0001 0002 ... .9996 9997 9999.
So, for (10f - 1)-2, count fromzero to IOOn
- 2 using n digitsfor each counting
number(and zero) and thenadd one to theresult.This is the block whichrepeats.
Here is a way to multiply(Ion _ l)-2 by any natural number less than lon.
Write:0 1 2 3 4 5 . . . (Ion - 1) but use n digitsforeach count,supplyingzeroes
whennecessary.To multiply(10n 1)2 by m, subtractm + 1 from10nthencircle
-

thisnumber(all n digits).
(a) Writethe firstuncirclednumber(all n digits).
(b) Count thenextm uncirclednumbers(n digitsper count),writethenumberon
whichyou land.
(c) Repeat (b), returning to the beginningwhen necessary(in a cyclicmanner).
(d) Stop when the numberimmediatelyfollowingthe circlednumberappears.
When you have finishedwriting, you have the block whichrepeats.
Examples:
7 x (10 - j)-2 Write:0 1(2)34 5 6 7 8 9 10-(7+ 1)=2
(a) 0
(b) 08
(c) 086
(c) 0864
(c) 08641
(c)(c)(c)(d) 086419753
7 x (l0-1)-2= 7x. 012345679=. 086419753 -Neat,huh?
33 x (102 _ 1)-2 Write(or think):00 01 02 03 96 97 98 99 Circle 66
(a) 00
(b) 00 33
(c) & (d) 00 33 67
33 x (102 - I)-2= 33 x. 00010203 .969799 =. 003367

14

This content downloaded from 131.91.169.193 on Tue, 27 Oct 2015 06:22:29 UTC
All use subject to JSTOR Terms and Conditions
Now let's multiply23.012 x.01235.
23.0 12 x.012 35 = (230. 12)(12. 35)(10-4).
And (230. 12)(12. 35) = (230)(12) + (230)(. 35) + (12)(. 12) + ( 12)(. 35)
= (230)(12) + (230)(35)(. 01) + (12)(12)(. 01) + (12)(35)(. 01)2

= 2760 + (8050)(. 01) + (144)(. 01) + (420)(. 01)2.

8050 x.01 = 080.80


+ 00.00
+ 0.50
+ .00
= 81.30
+ .01 1 was carried
= 81.31

144 x.01= 01.01


0.40
.04
1.45
420x(.01)2=10X42X(.01)2
=lOx.00428527... 1558
=(10 x.004285277012549739822467095194367921640648913376186103458830731558)
So 23.012 x .01235=(2760 + 81.31 + 1.45 +.0428527 ... 15580)(10-4)
=.28428105 . . . 56

The last partis not easy but it's merelyan additionproblem!!!!


Somethingwhichis possiblyworthmentioning:the algorithmspresenteddo not
depend on the base of the numerationsystem.For example: The algorithmfor
finding(10" - I)-2 can be expressedas
I on _2

ion(lOn-2) E i * 10-in +
i=O
(10 n-1)n(Ion-)2 1

It can be shownthattheformulais trueif we replace 10 withany naturalnumber


greaterthanone. Any readersinterestedin theproofof thisare invitedto writeme.
I would be happyto supplyit.
I would not be so happy to supplya proofof the algorithmfor multiplying a
naturalnumberby (Ion - I)-2. I have proved it to myselfbut the proofis in no
way elegant.This algorithmtoo can be used withany base numerationsystem.
A problemwhich I find extremelychallenging(if not defeating)is: Given a
repeatingdecimal,how can we find its multiplicativeinverse?You may wish to
workon that.

15

This content downloaded from 131.91.169.193 on Tue, 27 Oct 2015 06:22:29 UTC
All use subject to JSTOR Terms and Conditions