a”, a,

a

8

1. The iudices or exponents of a series of numbers is geoinetrical progression, proceeding from 1, are also called the logarithms of the numbers in that series.* Thus if a denote any number, and the geometrical series, 1, a', a’,a', r*, &c. be produced by actual multiplication, then 1, 2, 3, 4, &c. are called the logarithms of the first, second, third, and fourth powers of a respectively. Consequently, if, in the above, a be equal to the number 2, then I is the logarithm of 2, 2 is the logarithm of 4, 3 is the logarithm of 8, 4 is the logarithm of 16, &c. But if a be equal to 10, then I is the logarithm of 10,2 is the logarithm of 100, 3 is the logarithm of 1000, 4 is the logarithm of 10000, &c. The

11 series may be continued both ways from 1. Thus

a“, a, 1 1

1, a, a’, a, a, &c. constitute a series in geometrical progression, and, agreeable to the established notation in algebra, the indices, or logarithms, are -4,-3, -2, -1,0, 1, 2, 3, 4,&c. If a be equal to the number 2, then

I - 4 is the logarithm of - 3 is the logarithm of

16 - 2 is the logarithm of

O is the logarithm of 1, 1 is the logarithm of 2, &c. If a be equal to 10, then – 4 is the logarithm of - 3 is the lega

10000' 1 rithm of -,- 2 is the logarithm of -l is the lo 1000

100' garithm of o is the logarithm of 1, and 1 is the loga

10 rithm of 10, &c.

2. From the above it is evident, that the logarithms of a series of numbers in geometrical progression, constitute a series of numbers in arithmetical progression. Beginning with 1, and proceeding towards the right hand, the terms in the geometrical series are produced by multiplication, but their corresponding logarithms are produced by addition.-On the contrary, beginning with 1, and proceeding towards the left hand, the terms in the geometrical pro

1 is the logarithm of 1

> 4

1

gression are produced by division, but their corresponding logarithms are produced by subtraction. 3. The same observations apply to logarithms when they

1, 1, are fractions. Thus if a ñdenote any number, I

an 1 2

4 1, a”, ar, an, an, &c. constitute a series of numbers

a

2

3

1

1) un

3

2 1 in geometrical progression, of which

n, ng N, 0, 1, 2, 3, 4, &c. are the logarithms; and it is evident that

n,

3

-> n

n

the assertions in the last article hold true, both with respect to the numbers in geometrical progression and their corresponding logarithms. As a and n may be taken at pleasure, it follows that numbers in very different geometrical progressions may have the same logarithms; and that the same series of numbers in geometrical progression may have different series of logarithms corresponding to them.

4. If a be an indefinitely small decimal fraction, and successive powers of 1 + a be raised, then the excess of any power of 1 + a above that immediately preceding it will be indefinitely small. Thus let a = 00000000001, and then I + ale=1.0000000000200000000001; and 1 +al

l 1•000000000030000000000300000000001; and proceeding by actual multiplication to obtain higher powers of 1.00000000001, it will be found that the difference between two suuccssive powers is very small. If instead of supposing, as above, that a=00000000001, we suppose it only one millionth part of this value, then the successive powers of 1 + a will differ from one another by much smaller decimal fractions.

3. If therefore a be indefinitely small and successive powers of 1 + a be raised, a series of numbers in geometrical progression will be produced, of which the common numbers 2, 3, 4, 5, &c. will become terms. multiplication by 1 +a, an indefinitely small addition is made to the power multiplied, and by this indefinitely small addition, the next higher power is produced. Some power of 1 + a will therefore be equal to the number 2, or so nearly equal to it that they may be considered as equal. Continuing the advancement of the powers of 1 + a, the numbers 3, 4, 5, &c, for the same reasons, will fall into the eries.

6. The sum of the logarithms of any two numbers is equal

For on every

n

m

1 + a

1 + am

to the logarithm of the product of the same two numbers. Thus if 1 + a raised to the nth power be equal to the number N, and if I + a raised to the mth power be equal to the number M, then, by the preceding articles, n is the logarithm of 1 + an or of its equal N, and for the same reason m is the logarithm of M. Hence it follows that n + m= the logarithm of N ~ M, for N *M=i+a" x 1 + = 1 + a)" + m by the nature of indices. If the logarithim of N be subtracted from the logarithm of M, the difference is equal to the logarithm of the quotient which arises from

M am the division of M by N. For

N

=Italm-n, by the nature of indices. The addition of logarithms, therefore, answers to the multiplication of the natural numbers to which they belong; and the subtraction of logarithms answers to the division by the natural numbers to which they belong.

7. If the logarithms of a series of natural numbers be all multiplied by the same number, the several products will have the last-mentioned properties of logarithms. Thus if the indices of all the powers of 1 ta be multiplied by l, then using the notation stated in the last article, the logarithm of N is nl, and the logarithm of Mis ml, and the logarithmrof Nx M is nl + ml; for NxM=i+ani x 1 + almi = 1 + a)nl + mi, by the nature of indices. Also ml

M M 1+almi the logarithm of for

N'

N
Tam=1+aml_ni

. Hence

n the products arising from the multiplication of l into the indices of the powers of 1 + a, are termed logarithms, as are also all numbers which have the properties stated at the end of articles 6. It is on account of these properties that logarithms are so very useful in calculations of the highest importance.

8. If the indices of the powers of 1 + a, be multiplied by'a, the products are called the hyperbolic logarithms of the numbers equal to the powers of 1 ta. Thus if the

+ number N be equal to 1 + al», then na is the hyperbolic logarithm of N; and if the number M be equal to 1 + alm then ma is the hyperbolic logarithm of M. Hyperbolic logarithms are not those in common use, but they can be calculated with less labour than any other kind, and common logarithms are obtained from them.

- nl =

+

1

2

3

n

1

m

Xx3

2

2

3

9. If successive powers of a very small fraction be raised, they will successively be less and less in value. This truth appears most evident by putting the value in the form of a

1 vulgar fraction. Thus

100000 10000000000 100000 1

&c. 1000000000000000

10. Let it be required to determine the hyperbolic logarithm L, of any number N. Using the same notation as in the preceding articles, 17a"=N, and by extracting the nth root of each side of the equation, 1 ta=Nn. Put m=1

+

n, and 1 +x=N, and then Nn=1 + 2)" = (by the binomial

m 1

m-1

2 theorem) 1+mx+mX ха” + т Х Х + &c. =l+a. Now as a is indefinitely small, the power

lta of i+a, which is equal to the number N, must be indefinitely high ; or, which is the same thing, n must be indefinitely great. Consequently m must be indefinitely small, and therefore may be rejected from the expressions m-1, m-2, m-3, &c. Hence 1 being taken from each side of

mx2 mx3

mixt the above equation, we have a =mx

2

+

4 m.rs

&c. Each side of this equation being divided by m, 5

x2 x3 we have

--+ +

&c. But m= 1, and 3

5 x? 3

.x5 therefore

+ -&c,= L, the 2 3

5 hyperbolic logarithm of N, by article 8. This series, however, if x be a whole number, does not converge.

Let M be a whole number, and M=1, and then x is less than 1. For, multiplying both sides of the equation by 1-x, we have M- Mx=1, and therefore 1 -Ň=x. Now

M

=".

+

3

a

1

2

m

4

n

Int= 17.xlp. Then we have 1 +a=i-zle=

=

1 XP 1 - - =1-31' by putting r= - ) = - ræ+1 *

1

1-2

3

3

5

.22

2

3

2 x ******
*.**-***
x

x 23+ &c. But for the same reasons as above, r must be indefinitely small, and therefore may be rejected from the factors 1 - 1,5–2,7-3, &c.

r 1 Consequently, taking 1 from each side of the above equa

ra2 rx2 rr 1.25 tion, a=-rx

&c. But-s=

.;-$ and thererefore, dividing the left hand side of the equation

23 by-, and the other by-, we have ap =x+ ++;

xs + + &c. =the hyperbolic logarithin of M. 11. As, by the last article, the hyperbolic logarithm of N

x3 x4 x5 x x or 1 + x is a +

+ -&c. and as the 22

r hyperbolic logarithm of M or List + +

25 xo x? +

+ &c, the hyperbolic logarithm of NM, ita.

is equal to the sum of these two series, that is,

5

2

3

6

7

+

1

2

3

+6

+

7

273 2ars 2x7 equal to 2x+5+5++ &c. This series converges

++ faster than either of the preceding, and its value may be

x? expressed thus : 2 x(x++ + in+1

2n+2 12. The logarithm of = 2 x logarithm of +

2n + 1 the logarithm of

2n+112

For as the addition of lo

2n+1)-1 garithms answers to the multiplication of the numbers to which they belong, the logarithm of the square of any num

, ber, is the logarithm of the number multiplied by 2. Hence

2n+2 is 2 x logarithm of

But 2n+112

2n+1 2n+22 2n+1) 2n+22

492 +8+4 X 2n+1)2 2n+112-1 2n+D-1

496+41 n? +2n+1

- 1+1+1 n+1 n+n mXn+1

the logarithm of 2n+27

.

2

ent

n