XXXII .
Of the Theory of Circulating Decimal Fractions .
By John Robertson , Lib. R. S. THE great advantages arising from the use of Arthmetic , particularly in philosophy and commerce , is sufficiently known ; therefore , every step taken towards its perfection , has always been countenanced by those who were best acquainted with its nature and value .
On these motives I have been induced to offer the annexed paper to this learned Society .
Read Nov. 24 , 1768 .
Regiomontanus , it is said , first among Europeans , added to the then known arthmetic , an operation by decimal fractions ; which he exemplified in his triangular table .
Its utility was readily seen , and embraced in many nations , and particularly in this ; where it appears to have been cultivated in its theory , and facile modes of operation , more than in other places .
Many writers have remarked its excellency in numeral computation , and have pointed out compendiums to avoid the trouble of writing down superfluous figures ; particularly in the operations with concrete numbers , or those relative to money , weight , and measure ; where the gradations from one denomination to another do not proceed in an uniform progression .
In finding the decimal values of the fractional parts of concrete , and other numbers , it often hapD d2 pens , pens , that those decimals do not terminate , or end , with a few figures only ; and sometimes are infinite , or never end ; and among these are many which have one or more figures constantly recurring ; as in the following proportions , viz. 3 : 2 : : 1,0000 , &amp;c. : 0,6666 , &amp;c. and 12 : 5 : : 1,0000 , &amp;c. : 0,4166 , &amp;c. also 7 : 3 : : 1,0000 , &amp;c. : 0,428571,428571 , &amp;c. In operations , with such recurring decimal fractions , particularly in multiplication and division , the work will either be longer than necessary , or be very inaccurate , if the numbers are not considered as circulating ones : and to come at the true results of such operations , several authors have given precise rules ; and some of them have shown the principles upon which those rules were founded .
In the annexed paper those principles are , endeavoured to be , exhibited in a different , and in a more general and concise manner , than has hitherto been shown : but the modes of working are not here annexed , as they are to be found in Cunn , Malcolm , Marsh , and others ; and may hereafter be fully exemplified in a treatise of Arithmetic , by the author hereof , considered in a more mathematical order , than what has hitherto been appropriated to this most useful science .
GENERALPRINCIPLE S. 1 .
Number is supposed to begin at unity , and from thence to ascend and descend : those terms ascending above unity , are integers ; and those descending below unity , are fractions .
When in the ascending and descending parts of the scale , the gradation proceeds by a tenfold value from the right hand towards the left , the rank of numbers , thus generated , is called the decimal scale .
As every place in this scale is ten times the value of its next right hand place ; therefore the first place in thefractional part , is 1/ 1 0 of the place of units ; and the second , third , fourth , &amp;c. descending places in the fractional part , is 1/ 1 0 0 , 1/ 1 00 0 , 1/ 1 000 0 , &amp;c. part of the place of units .
Therefore every decimal fraction is equal to a series arising from m ultiplying the first , second , third , fourth , &amp;c. terms of the decreasing geometrical progression 1/ 1 0+ 1/ 1 00+ 1/ 1 000+ 1/ 1 000 0 , &amp;c. by the first , second , third , and fourth , &amp;c. terms in the given fraction respectively .
Thus .
Let the given fraction be 0,3587 or 358 7/ 1 000 0 .
Then 0,3587=1/ 1 0X3+ 1/ 1 00X5+ 1/ 1 000X8+ 1/ 1 0000X7 = 3/ 1 0+ 5/ 1 00+ 8/ 1 000+ 7/ 1 0000 =3 00 0/ 1 0000+50 0/ 1 0000+8 0/ 1 0000+ 7/ 1 0000=358 7/ 1 000 0 .
2 .
Every decimal fraction arises from division , when the dividend is less than the divisor .
For , divisor : dividend : : 10 : first term of the fraction ; : : 100 : sum of the first and second terms ; : : 1000 : sum of the first , second , and third , terms , &amp;c. And according to the ratio of the divisor to the dividend , the quotient , or decimal fraction , will be finite or infinite .
3 .
Among those decimal fractions which are infinite , or do not end , some of them recur , or circulate ; that is , the same figure or figures run over again and again ad infinitum .
As 0,333 &amp;c. ; 0,2323 &amp;c. ; 0,758758 &amp;c. ; 0,999 &amp;c. Here 0,333 &amp;c. =3/ 1 0+ 3/ 1 00+ 3/ 1 000+ 3/ 1 000 0 , &amp;c. 0,2323 &amp;c. =2/ 1 0+ 3/ 1 00+ 2/ 1 000+ 3/ 1 000 0 , &amp;c. 0,785785 &amp;c. =7/ 1 0+ 8/ 1 00+ 5/ 1 000+ 7/ 1 0000+ 8/ 1 00000+ 5/ 1 00000 0 , &amp;c. 0,999 , &amp;c. =9/ 1 0+ 9/ 1 00+ 9/ 1 000+ 9/ 1 000 0 , &amp;c. The circulating fraction 0,999 , &amp;c. is equal to 1,0 .
For the difference between 1,0 and 0,999 &amp;c. is less than can be assigned .
To save a repetition of figures , it is usual to mark the first and last of circulating expressions , with points over the figures .
Thus , 0,333 &amp;c. is wrote 0,3 0,2323 &amp;c. 0,23 0,785785 &amp;c. 0,785 4 .
Like circulating decimal fractions are those which have each the same number of circulating places ; and begin to recur each at the same name .
Every finite decimal fraction may be considered as infinite ; ciphers being used as the circulating part .
Either place of a circulating expression may be taken as the first ; observing that the number and order of the circulating places be not altered .
For as the decimal fraction arises by division ; if either place of the recurring figures be taken for the first , the others will from thence regularly circulate .
Hence several unlike circulating decimal fractions may be made to begin and end at places of like names .
5 .
If in the decimal scale 10 , 100 , 1000 , 10000 , 100000 , 1000000 , &amp;c. continued indefinitely , be selected any rank of equi-distant terms , such , that whatever term therein is taken for the first term , and the first term is made the common ratio to the rest ; then will the sum of the reciprocals of those terms , be equal to the reciprocal of the number which is unity less than the first term .
Thus 1/ 10 + 1/ 100 + 1/ 1000 + &amp;c. = 1/ 9 ; 10 being the 1 st term .
1/ 100 + 1/ 10000 + 1/ 1000000 + &amp;c. = 1/ 99 ; 100 being the 1 st term .
1/ 1000 + 1/ 1000000 + 1/ 1000000000 + &amp;c. = 1/ 999 ; 1000 being the 1 st term .
For 1/ 10 + 1/ 100 + 1/ 1000 + &amp;c. = 100 + 10 + 1/ 1000 = 111 &amp;c. / 1000 &amp;c. 1/ 100 + 1/ 1000 + 1/ 1000000 + &amp;c. = 10000 + 100 + ~/ 1000000 = 10101 &amp;c. / 1000000 &amp;c. 1/ 1000 + 1/ 1000000 + 1/ 1000000000 + &amp;c. = 1000000 + 1000 + 1/ 1000000000 = 1001001/ 1000000000 &amp;c. But 111 &amp;c. ~9= 1000 &amp;c. ( by 3d ) Then 111 &amp;c. / 1000 &amp;c. = 1/ 9 10101 &amp;c. ~ 99 = 100000 &amp;c. Then 10101 &amp;c. / 1000000 &amp;c. = 1/ 99 1001001 &amp;c. ~ 999 = 1000000000 &amp;c. Then 1001001 &amp;c. / 1000000000 &amp;c. = 1/ 999 Hence the reciprocal of a number consisting of n places of 9~s , is equal to a circulating number of n places , the right hand figure being 1 , and the rest 0~s .
Thus , 1/ 9 = 111 &amp;c. / 1000 &amp;c. = 0 , ~1 1/ 99 = 10101 &amp;c. / 1000000 &amp;c. = 0 , ~0~1 1/ 999 = 1001001 &amp;c. / 1000000000 &amp;c. = 0 , ~00~1 1/ 9999 = 10001001 &amp;c. / 1000000000000 &amp;c. = 0 , ~00~0~1 6 .
If the reciprocal of a number consisting of n places of 9~s , be multiplied by a number D , not exceeding n places ; the product will be a circulating decimal fraction of n places , the right hand ones being the same digits as are in the number D. Let D= 3 ; or D= 23 ; or D= 785 ; or to any other number .
Now 1/ 9 = 111 &amp;c. / 1000 &amp;c. ; 1/ 99 = 10101 &amp;c. / 1000000 &amp;c. ; 1/ 999 = 1001001 &amp;c. / 1000000000 &amp;c. ( by 5th ) .
Therefore 1/ 9 , or 111 &amp;c. / 1000 &amp;c. ~3= 333 &amp;c. / 1000 &amp;c. = 0 , ~3 ~3= 30303 &amp;c. / 1000000 &amp;c. = 0 , ~0~3 1/ 99 , or 10101 &amp;c. / 1000000 &amp;c. ~ 23 = 232323 &amp;c. / 1000000 &amp;c. = 0 , ~2~3 ~3= 3003003 &amp;c. / 1000000000 &amp;c. = 0 , ~00~3 1/ 999 , or 1001001 &amp;c. / 1000000000 &amp;c. ~ 23 = 23023023 &amp;c. / 1000000000 &amp;c. = 0 , ~02~3 ~ 785 = 785785785 &amp;c. / 1000000000 &amp;c. = 0 , ~78~5 7 .
Hence every circulating decimal fraction will be equivalent to a vulgar fraction , wherein the numerator is those circulating figures , and the denominator consists of as many 9~s , as are figures in the numerator .
Thus 0~3 = 3/ 9 .
For 1/ 9 ~3= 3/ 9 = 0 , ~3 ( by 6th ) 0 , ~0~3 = 3/ 99 1/ 99 ~3= 3/ 99 = 0 , ~0~3 0 , ~2~3 = 23/ 99 1/ 99 ~ ~2~3 = 23/ 99 = 0 , ~2~3 8 .
Hence a circulating decimal fraction , of any number of places , being multiplied by a number of as many 9~s , will give a finite expression , having the same figures as are in the circulating one .
Thus 0 , ~6 ~9= 6 .
For 9 : 1 : : 6 : 0 , ~6 0 , ~0~6 ~ 99 = 6 .
99 : 1 : : 6 : 0 , ~0~6 0 , ~2~5 ~ 99 = 25 .
99 : 1 : : 25 : 0 , ~2~5 0 , ~62~5 ~ 999 = 625 .
999 : 1 : : 625 : 0 , ~62~5 Hence it appears , that , in common multiplication , the product of a circulating number , by its proper denominator , in 9~s , will be deficient of the true product by that circulating number .
Thus ~0 , ~6 ~9= 5,4 ; then 5,4 +,6 = 6 .
For 6/ 9 = 0 , ~6 0,0~6 ~ 99 = 5,94 ; then 5,94 + 0,06 = 6 .
For 6/ 99 = 0,0~6 0 , ~62~5 ~ 999 = 624,375 ; then 624,375 + 0,625 = 625,0 Hence .
Any finite number is in proportion to the same number recurring , as the proper denominator of the circulate is to that denominator increased by unity .
Thus 9 : 10 : : 6 : ~6 .
For ~6 ~9=6~ 10 99 : 100 : : 25 : ~2~5 .
For ~2~5 ~ 99 = 25 ~ 100 .
SCHOLIU M. If to the preceding articles , be joined the compendiums of multiplying and dividing by any number of 9~s , they will constitute the whole of the theory , upon which depend all the operations with circulating numbers : for as these have 9~s for their denominator , wanting unity in the lowest place to make them 10~s ; therefore unity for every 9 is applied in some additions and multiplications : Or , the circulating parts being reduced to finite number ; then working with them by the common rules , will give finite results ; which results are to be reduced to circulates by contrary operations to what were used to reduce the circulates to finites .

