0 ratings 0% found this document useful (0 votes) 354 views 82 pages An Easy Introduction To The Slide Rule - IsaacAsimov-1965
out of print book by asimov on how to use a slide ruler
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content,
claim it here .
Available Formats
Download as PDF or read online on Scribd
Carousel Previous Carousel Next
Save An Easy Introduction to the Slide Rule - IsaacAsim... For Later A COMPUTER
IN YOUR POCKET..
In this book, Prote:
Asimov, noted
With his usual
enthusiasm and talent
for making things
both entertaining and
understandable,
Professor Asimov
teacher, a
explains the principles
“does! mas of the slide rule.
impressive as a glant Once you thoreughly
electronic computer, but understand these
ny advantages. prifciples—once you
mall enough know what you are
doing and why-it will be
Itneed not cost simple to use the slide
| more than a couple rule on problems that
of dollars, it can't arise from day to day.
Be eave ete “A delightful book for
and, best of all, learning.”
itcan solve almostany Sanna.
numerical problem that
you meet up with
under ordinary
clroumstances. To add
fo all that, it
ie simple to operate.
if you know grade
sehoo! arithmetic,
you ean uso a slide
tule, even though
you may not quite see
how it worke!"’
ithe
tt
to put in your pocket,
FAWCETT WORLD LIBRARY
om.
AOUISy Obes! —F1AY ICN SHEL NOLORGOULNEASTS NY
yoy sopwosg
Homey
A Fawcett Premier
BookMeck. (Mihov
ily 5, HqOther Faweett Premier Books by
Isaac Asimov:
Realm of Algebra
Realm of Measure
Realm of Numbers
Introduction
to
the
Slide
Rule
with Diagrams by
npn
Mouberof American
Fann Book Publishers Council
pene a TT
3 0
freuen ine oo
wagesh Fact Dremies book repented by wrangement vith
foughton Mifflin Company, Boston. This book contains the
complete text of the original hardcover edition.
Copyright © 1965 by Isaac Asimov. All rights reserved including
the right to reproduce this book or parts thereof in any form.
Library of Congress catalog card number: 65:19204
PRINTING HISTORY
First Houghton Miflin Co., printing, September 1965,
Second printing, June 1966
First Fawcett Premier printing, March 1967
Published by Faweett World Library,
67 West 44th Street, New York, N.Y. 10036
PAINTED IN THE UNITED STATES OF AMERICA
Contents
Arithmetic by Distance
The Addition Rule
Logarithms
Logarithms on Wood
Multiplication
Division
Powers
Roots
Index
9
39
5B
82
107
131,
152
159To a friend
who has accompanied
and served me
loyally for
twenty years
My Sume Rute
1
Arithmetic by Distance
To Begin With—
We nave all heard, these days, of electronic computers.
These marvelous instruments, which came into use during
World War II, are capable of performing in a few
seconds work that might take years if all we could use
were pen and paper.
‘There are times when arithmetical problems come our
way and we might wish that we ourselves owned such a
computer to do the work for us. Such a situation would
have its disadvantages, however. Electronic computers
are bulky, expensive, complicated, and can be handled
only by people with special training.
Besides, electronic computers aren't at their best
when used for everyday problems. That would be like
trying to shoot a fly with naval artillery.
For a fly, an ordinary swatter is much better, and for
ordinary mathematical problems, we could best use a
really simple computer.
‘There happens to be a simple computer, just suitable
for everyday computations, that was invented about
350 years ago. It isn't electronic; there are no electric
currents involved. In fact, it is no more than a piece of
wood with some marks on it. It looks like a ruler except
that it has a middle piece that can slide back and forth,
so that it is called a slide rule.
a eeeerC_T lsA slide rule doesn’t seem as impressive as a giant
electronic computer, but it has many advantages. It is
small enough to put in your pocket, it need not cost
more than a couple of dollars, it can’t go out of order, and,
best of all, it can solve almost any numerical problem
that you meet up with under ordinary circumstances.
‘To add to all that, it is simple to operate. Tf you know
grade-school arithmetic, you can use a slide rule, even
though you may not quite see why it works! If you have
taken some high-school mathematics also, the reason for
its success can be explained with very little trouble.
In this book, I shall start from the beginning and try
to show you both how and why it works.
To be sure, merely reading this book will not make you
an expert at handling the slide rule. For that you will
need practice and I am not including practice exercises
as part of this book.
However, once you thoroughly understand the prin-
ciples of the slide rule— once you know what you are
doing and why — then it will be simple to set yourself
problems, It will be simple for you to use the slide rule
on problems that arise from day to day.
As you practice you will become expert, and you will
be amazed to see how, by merely sliding one piece of
marked wood against another, computations that seem
very complicated can be completed in a few moments.
In fact, if the time comes when you have a job in
which numerical computations have to be carried
through frequently, you will want your slide rule with
you at all times, Without it, you would feel like a doctor
without his stethoscope or a painter without his brush.
Let's consider, then, how a piece of marked wood can
help us in our calculations.
One Ruler
One of the earliest tools used by civilized men engaged
in fine and accurate work — as in making plans for build-
ing temples or tombs — must have been a piece of wood
or ivory or metal which had a straight edge. By sliding
a pencil, or any marking device, down this straight edge,
a straight line could be drawn.
You understand how important it is, in drawing up
any accurate plan, to produce perfectly straight lines,
but to do so some guide is absolutely necessary. Try
drawing straight lines freehand, that is, without using
some straight edge to guide you, and see what a sloppy
appearance it makes.
‘An instrument possessing a straight edge and nothing
more is called just that, a straightedge.
It is simple, however, to put the straightedge to
another use, and make it a means for determining the
length of a straight line.
For this purpose it is only necessary to make small
marks upon it —marks spaced a fixed difference apart.
Such a marked straightedge is a rule or a ruler.
The ruler with which we are most familiar has the
marks upon it spaced an inch apart, and is 12 inches long
all together. Since 12 inches make 1 foot, such a ruler is a
foot-rule.
The foot-rule has its inches subdivided into halves
and quarters, usually eighths as well, and sometimes
even into sixteenths (see Figure 1).
Such a ruler can be used to determine the length of awe
Figure |
line or to draw a line of a particular length, and we are
all familiar with its use.
Can we, however, use it for some other purpose than
measuring length? Suppose we think about that a bit . . .
When we say that 1 + 1 = 2, we are talking about
“pure numbers.” The 1’s and the 2 in that sum do not
represent one of anything, or two of anything. Still the
sum can be made to apply to numbers that are not pure
— to count objects.
For instance, 1 apple +1 apple =2 apples; 1 chair
+1 chair =2 chairs; 1 star-+1 star =2 stars; and 1
inch + 1 inch =2 inches. We can say this not only for a
sum like 1-++1=2, but for any arithmetical problem
that involves addition or subtraction. We know that
72+ 28= 100 and that 35 —20= 15, and we can be
sure, therefore, that 72 inches + 28 inches = 100 inches,
and that 35 inches — 20 inches = 15 inches.
In other words, if we add inches to inches, or subtract
inches from inches, the numerical portion of the answer
we get would be the same as the numerical portion we
would have obtained if we had used apples, chairs, stars,
or anything else. It would be the same, indeed, as
answers we would have gotten in using pure numbers.
10
‘This means that if we could somehow use some device
to prove that 45 inches and $2 inches taken together
make up a length of 77 inches, then we can be sure that
45 + 82=77. The device we would use to put lengths
together in this manner would, in effect, do our addition
for us.
Let's take a very simple example. Look at a ruler (or,
if you don't have one handy, look at Figure 1) and,
starting at the left edge, count off 2 inches with your
s finger. Your finger is now pointing to the number 2.
From there, count off 3 more inches. Now your finger is
pointing to the 5. What you have done is shown yourself
that 2 inches + 3 inches = 5 inches, and you can deduce
from that that 2+ 3 = 5.
‘Try again. Count off 3 inches and, starting from the
place you reached, count off 5 more, and you will find
that +5=8. In the same way you can use an ordinary
foot-rule to prove to yourself that 4 +7 = 11,1+8=9,
or even that 6 + 3 = 3 + 6, for both these sums come
out to 9
Subtraction works just as well. Count off 10 inches
and you are at the number 10, If you want to subtract 3
inches from that, count them backward, chopping each
inch off the total length. When you finish you find yous
self at the number 7. This means 10 inches — 3 inches =
Tinches, or 10-3
What we are doing is manipulating lengths and
using those lengths to tell us something about numbers,
Lengths are not numbers, of course, but in certain ways
they follow the same rules that numbers do. Lengths
have properties that are analogous to those of numbers in
some ways. If we then use lengths to work out, or
"“compute,” the answers to number problems, we are
using the foot rule as an analog computer.
‘An example of a very elaborate analog computer is
Univac, which is used by television studios to work out
the course of elections in progress and to predict the
winner. (Univac stands for Universal Analog Computer.)
Univae makes use of flashing electric currents that
alter the position of many thousands of tiny switches in
ionths of a second. The properties of the changing
switch positions are analogous to those of numbers, which
is why Univac can be used to solve problems, However,
an ordinary ruler, although much, much simpler than
Univac, can also be used as an analog computer, and
we've just done it. In fact, it is easier to determine that
2+8=5 ona foot-rule than it is on Univac.
Two Rulers
Next, let’s see how we can make a ruler more useful
for this new purpose — not of measuring lengths, but of
helping us work out simple additions and subtractions.
In the first place, a 12-inch rule is inconvenient. It only
exists because 12 inches make 1 foot, but we are not inter-
ested in that just now. We are interested in arithmetic
and in arithmetic we base everything on the number 10.
Suppose you start with nothing (zero) and begin to
count: 0, 1,2, 3, 4, 5, 6, 7,8, 9. Counting the zero, there
are ten different symbols (or digits) for the first ten
numbers. The next number, 10, makes use of two digits,
with a 1 on the left. The 1 on the left is kept there in
the next few numbers while the digit on the right goes
through the same series over again: 10, 11, 12, 13, 14, 15,
16, 17, 18, 19, With a 2 on the left, we again go through
2
the same series; then we do the same with a 8 on the left,
with a 4 on the left and so on,”
If, therefore, we learn to work with the numbers from
0109 to begin with, it will be easy to apply the knowledge
to the higher numbers which, in a way, are merely repeti-
tions of the first set. Consequently, a 10-inch rule will
be sufficient.
Let's use our 10-inch rule to deal with the addition
problem 2 +3 = 5. To begin with, we count off 2 inches
from the left, but we don’t really have to do any actual
counting, If we count off 1 inch from the left, we end at
the figure 1; if we count off 2, we end at 2; if we count
off 8, we end at 3 and so on. Therefore, for the first
number in the problem, 2 + 8 = 5, we move straight to 2,
without counting.
Starting at 2, we next have to count off 3, but why
count it off? Since we are actually dealing with distances
rather than with numbers, why not measure the 3 inches
with another ruler?
Imagine a second ruler, with the inches marked off
at the bottom, rather than at the top. Imagine the second
ruler, edge to edge, on top of the first. The 1’s will match
on the two rulers, and so the 2's, the 3's, and so on
(Figure 2).
Slide the top ruler to the right until its left end is over
the 2 on the bottom ruler. Now, you see, the top raler
is measuring off the inches starting at the bottom 2
(Figure 3). If you look at the 3 on the top ruler, that
marks a distance of 3 inches and immediately mderneath
it is the 5 of the bottom ruler. This is the method of
* For a fairly detailed discussion of how our number system
works, you might refer to another book in this series: Realm of
Nembere
123showing that 2+ 8 =5 by measuring distances.”
If you keep the top ruler in the position shown in
Figure 8, with its left end over the bottom 2, then
you can see that 2+1=, because the top 1 is over
the bottom 8, and, for similar reasons, 2+2= 4; and
2+4=6, The single position of the ruler gives us the
answers to a whole family of sums. It is easy to see how
this system can be used to give answers to still other
sums: 3+5=8; 6+1=7, and so on. .
When two rulers are placed edge to edge, with one
sliding along the other, the result is a slide rule. (You
can also think of it as two sticks, one slipping along
another, People who use a slide rule often, sometimes
call it a “slipstick,” but this is considered slang. )
A slide rule consisting merely of two ordinary rulers
g but convenient, however. The ordinary ruler
‘Tt may seem silly to you to use so roundabout 1 method for
Snding out something as simple as 2+ 3 ~ 5, Right now, how-
fever, we are only working out the technique. Things will ree
main very simple for quite a few pages, but before long we will
be tackling more dificult problems —and solving them with
just as litle trouble as 2 + 3 = 5 gives us now.
Figure 2
igure 3
tapers to a narrow edge at the inch markings, in order
to make it easier to draw a straight line and measure its
length. Sometimes, the narrow edge even has a thin
metal strip down its length to make it still easier to
draw lines,
We don't use a slide rule, however, for drawing
straight lines. We want to design it, instead, in such a
way as to make it easy to slide one ruler against the
other, With thin edges, it is difficult to do this, and with
metal strips it is practically impossible. It would be
better to use two blunt rulers that are of even thickness
all the way across, and that meet, therefore, at thick
edges. (In fact, slide rules are usually a quarter of an
inch thick.)
Even meeting at a thick edge has its problems. If the
edges are quite smooth, the top ruler can easily slide
sideways off the bottom one. For that reason, the bottom
ruler is generally made with a groove down its length
and the top ruler has a small tongue of wood that fitsinto the groove, Now the top ruler can slide back and
forth easily, without any danger of slipping off sideways.
‘To be sure, the top ruler can still move upward easily.
And if the slide rule happens to be turned upside down
at any time, the top ruler, tongue and all, will fall out
of the groove.
To prevent tipping, the top ruler can be made with
a tongue of wood on top as well at on the bottom, and
the upper tongue can be made to fit into a third ruler
with a groove. What was the top ruler becomes a middle
ruler. The next step is to bolt the uppermost ruler to
the lowermost ruler, holding them firmly in place, while
the now middle ruler can slide back and forth easily
between them (Figure 4). The middle ruler cannot
tip, and the slide rule cannot fall apart.
You might suppose that if you hold such a slide rule
by one end and let it dangle, the middle ruler will simply
slip out. However, the fit is usually made tight enough
so that friction will keep it in place. To move that
middle ruler, you actually have to push it.
Making the Rulers
Ordinarily, you could follow the directions I will give
you for using a slide rule by following them on an actual
instrument. That would serve as an additional guide to
that of the illustrations that will be presented.
For a while, however, we are going to concern our-
selves only with addition and subtraction, operations
which are not conducted on the slide rules that are
actually manufactured. (‘This is for the very good reason
that addition and subtraction are too simple to require
a slide rule.) 6
Figure 4
In fact, in describing the use of rulers in performing
addition and subtraction, I hesitate to call the device
aslide rule lest you confuse it with the slide rules actually
manufactured for use in other operations. Let us there-
fore call the addition-and-subtraction device an addition
rule and use only that term for the purpose.
In order to follow the workings of the addition rule as I
describe them, you may have to rely very heavily on the
illustrations. You cannot be guided by the use of an
actual addition rule, unless you choose to make one for
yourself. Fortunately, it is not difficult to make an addi-
tion rule.
To do this, get a sheet of flexible cardboard, of the
type used in the manufacture of folders, and cut out two
pieces (A and B) according to the measurements shown
in Figure 5.
Fold Piece A along the lines a and b. The end pieces,
which will overlap, are pasted together with mending
tape, so that we now have something that looks like a
flattened cylinder with a long rectangular hole in it.
1ee
we»
Figure 5
Piece B will fit into Piece A snugly and can he moved
back and forth within it. Piece A will now serve as the
top and bottom ruler, “bolted” together with tape, while
Piece B will be the middle ruler. Where Piece A and
the lower portion of Piece B meet mark off inches from
© to 10 on both as shown in Figure 6.
If you cannot find a piece of cardboard 14 inches
Jong, you can make a sinaller addition rule by working
with measurements half the size of those indicated in
Figure 5 and then marking off the 0 to 10 divisions in
half-inch units.
The result, in either case, will be an addition rule
which, however crude, will give you the feel of the
instrument,
18
2
The Addition Rule
Beyond Ten
We nave reached the point now where we should
use proper terms, and no longer speak of top rulers,
bottom rulers, and middle rulers. The part of the ad
tion rule which slides back and forth is the slide. The
rest of it is the body.
We can identify the range of numbers on the slide
as the “S-numbers,” and the range of numbers on the
body as the “B-numbers.” If I speak of S-7 and B-5,
I mean the 7 marked on the slide and the 5 marked on
the body.
‘What was the left end of an ordinary ruler — an end
not marked by a number, as you can see if you look at
one — is now represented by the number 0 on the addi-
tion rule,
Instead of saying, then, “Move the middle ruler to
the right until its Jeft end is over the 5 on the bottom
ruler,” I will say, much more simply, “Place $-0 over B-5.”
In order to add 5 and 2 on the addition rule, the
directions would be: Move S-0 over B-5, and under-
neath $-2, you will find B-7 (Figure 7), indicating that
5+2=7.
By now, though, you have undoubtedly realized that
19Figure 7 ua
our addition rule will give us the answer to only the very
simplest addition problems. Suppose we wanted the
answer to 7+5, a problem which is still quite simple
and which we know has the answer 12. Yet, at first
glance, our addition rule can’t help us.
If we place $-O over B-7, then we would expect to
find the answer under S-3. The trouble is, however, that
$-5 is not making contact with the B-numbers at all
(Figure 8), Now what?
One solution is to make the body of the addition rule
longer, 20 nits long, rather than 10. That would mean
we can mark off numbers from 0 to 20. Let us remember,
however, that the numbers 10 and beyond are repetitions,
at least in their right-hand digits, of the numbers under
10. We can make this plainer if, in the numbers from
10 to 20, we place the digit to the left in parentheses,
so that we can concentrate on the digit to the right
(Figure 9).
Using this extended addition rule, we can add 7 and
5 by placing $-0 over B-7 and finding $-5 over B-(1)2.
This shows us (Figure 10) that 7+5=12.
‘The trouble is that the longer an addition rule is, the
cob ye
Figure 8
less convenient it is. No one would want to use one
in which the bottom half of the body stuck out like an
elephant's tusk. Is there any other way out?
To find one, look at Figure 10 again. Notice that -0
is over B-7 and that S-(1)0 is over B-(1)7. Ignore the
parentheses and you'll see that the situation has re-
peated itself. What's more, this would happen every
time. Hf S-0 were over B-5, then S-(1)0 would be over
B-(1)5; if $-0 were over B-2, then S-(1)0 would be over
B-(1)2, and so on.
There's no puzzle about why this should be. In
moving from $-0 to S-(1)0, we are adding 10 and we
know that 2+ 10= 12, 5+ 10= 15, 7+10=17 and so
on. The right-hand digit of 10 is 0, and when we add
10 to any number, the right-hand digit of that number
(to which we are adding zero) naturally remains un-
changed.
In that case, it doesn’t really matter whether we place
$-0 or S-(1)0 over a B-number. In either case, the right-
hand digit of the answer will be the same, and we must
only remember to adjust the left digit. If we use S-0,
the left digit of the answer is 0 and is omitted. If we use
Figure 9Figure 10
T4002
S-10, the left digit of the answer is 1.
For instance, let’s go back to our ordinary 10-unit
addition rule and consider 7+5 again. This time, in-
stead of putting S-0 over B-7, we place S-10 over B-7
(Figure 11).
If you compare Figures 10 and 11, you will see how
they duplicate each other. In Figure 10, on the 20-unit
addition rule, $-4 is over B-(1)1, 8-6 is over B-(1)3, and
S-8 is over B-(1)5, while in Figure 11, on the 10-unit
addition rule, 8-4 is over B-1, $-6 is over B-3 and S-8 is
over B-5.
On the 10-unit addition rule, if we place S-10 over
some B-number, we need only add the left-hand digit 1
to our answer to get what we would have obtained with
the much less convenient 20-unit addition rule. In Figure
11, with $-10 over B-7, we find that S-5 is over B-2. We
add the left-hand 1 and we know that 7+ 5 = 12.
We can also tell by Figure 11 that 7+4=11,7+7=
14,7+9=16.
‘Thus, by making use of both $-0 and $-10, we can
carry our additions on the 10-unit addition rule up to a
sum of 20. And yet 20 isn’t the highest possible sum
there is either. What if we wanted to get the answer to
47+14?
In order to tackle additions of this sort, let's take a
detour in what seems the opposite direction — numbers
smaller than 1 rather than Boyer than 20.
Fractions
So far, our addition rule only has digits on it, and we
have been dealing only with a few small whole numbers.
But you know that there are fractions — numbers that are
smaller than one, or that are intermediate in value
between neighboring whole numbers. In fact, if you
Jook at an ordinary ruler again, you will find fractions
there, for the inches are divided into halves, quarters,
and eighths, and often sixteenths, too.
Suppose we divide each unit distance on our addi-
tion rule into halves, quarters, and eighths (Figure 12).
You've undoubtedly had practice reading a ruler, so
that you will probably have no difficulty in reading a
particular mark in Figure 12 as 3%, or 4%, or 7%. And,
of course, you will find the simple fraction % between
Gand 1.
Fractions work as well as whole numbers do on the
addition rule. Why shouldn't they? If we add 1% inches
Figuee ItFigure 12
% om m
and 2% inches, we would get 3% inches, and that tells
us 14+2%=8%. Try it on the addition rule. Place
S-0 over B-1% and $-2% is found to be over B-3%
(Figure 13).
By proper manipulation, you can find in the same
way that 2% + 4% =7%; that 54 +1%=7%, and so
‘on. (The addition rule should impress you a little more
now. It adds fractions as easily as it adds digits, whereas
with pen and paper, it is considerably harder to add frac-
tions than digits.)
For addition of fractions yielding sums over 10, we
use $-10, To add 5% and 7%, for instance, we place $-10
over B-5% and find S-7% over B-2% (Figure 14). We
insert that left-hand digit 1, and find that 5% +7% =
12%.
‘There is no question that this system of adding
fractions is very handy, but we can do still better. Re-
member that we changed the 12-inch ruler to a 10-inch
ruler because in our number system the right-hand digits
™ Figure 13
eH
vie EN
Figure 14
enn
repeat after intervals of 10 (see page12). This makes
10a particularly useful number in calculations of alll sorts.
Following this system of emphasizing 10, we can divide
each unit on the addition rule into tenths, instead of
into halves, quarters, and eighths,
The result is shown in Figure 15. The %e mark (which
is equal to %) is made longer than the others so that it
stands out. You may be less accustomed to this system
of subdividing units, but you should have no trouble
picking out 8%e, or 740, or 2%o.
It may seem at first that we have simply changed one
set of fractions for another and you may wonder what
we have gained. But tenths, you see, are special fractions
because of our 10-based number system. With tenths,
we can use decimal fractions. Thus 1%o can be written
1.1; 8%o can be written 3.6; 9%o can be written 9.9, and
so on. By using tenths, then, we have switched our ad-
dition rule to the decimal system.
‘We can use the addition rule as easily for decimals as
for ordinary fractions or for simple digits. If we want the
sum of 1.7 and 1.4, we place S-O over B-1.7 and find
S-14 over B-3.1 (Figure 16). We can tell, then, that
L7+14=3.1 It is just as easy to discover that 2.3 +
7.1=94, or that 5.8 + 16=7.4.
For sums over 10, we use.S-10. If we place $-10 overi
Figure 18
wide unit can be divided conveniently are 0.02 inch units,
Each tenth would have five subdivisions (Figure 18),
so that you could read a number like 8.12 or 8.16.
Using such an addition rule, you can find, without too
much difficulty, that 0.24 + 8.12 = 8.36 or that 8.12+
6.72 = 14.84. This will tell you that 2.4 + 81.2 = 83.6
and that 81.2 + 67.2 = 148.4. It will also tell you that
24 + 812 = 836 and that 812 + 672 = 1484.
It may seem to you that now we can only handle
numbers with right-hand digits that are even, for the
smallest divisions shown in Figure 18 represent 0.02, 0.04,
0.06 and 0.08. There is no division equivalent to 0.03,
for instance, so that a number such as 7.33 could not
he read directly.
‘And yet why not? You have no trouble finding 7.82
and 7.34 on the addition-rule as we now have it. Between
them is an empty space with no markings and we can
imagine a fine line splitting that empty space in two.
This imaginary line (shown in Figure 19 as a dotted
line) would be 7.33. You might even imagine lines drawn
closer to 7.32 than to 7.84, or vice versa and representing
7.825 or 7.385. These lines, which wouldn't be shown
directly on even the most cagefully manufactured 10-inch
Figure 19
addition rule (if any were manufactured at all) can
nevertheless be estimated.
To be sure, no matter what we do, there are limits to
how finely we can read subdivisions. Our estimates are
bound to be uncertain and we can’t push them too far.
Im addition, the most carefully made markings may be
very slightly off position and even if they aren't, the
markings have a perceptible thickness (they must have)
and ideally they should have no thickness at all.
We are, in short, condemned to inexactness. To see
why that is, let’s go back for a moment to ordinary
fractions.
Inexactness
The decimal addition rule can handle ordinary frac-
tions, too, provided those ordinary fractions are first con-
verted to decimals. We can add 5% and 2% without
trouble if we convert the numbers to 5.50 and. 2.75.
Using the addition rule, we find that 5.50 + 2.75 = 8.25.
If we wish, we can change 8.25 back to 84 and then we
have determined that 5% + 2% = 8%.
The conversion of fractions to decimals, however, can
present the addition rule gh more than it can reallyFigure 15
B-8.8 and find $-5.6 over B-3.9 (Figure 17), we need
only add the left-hand digit 1 to conclude that 8.3 +
5.6= 18.9.
Moving the Decimal Point
We are now ready to see a great advantage in the use
of decimals over ordinary fractions. The decimals make
it possible for us to handle additions that reach sums
well over 20.
‘We have already determined by means of the addition
rule that 1.7 + 1.4 = 8.1 (Figure 16). Actually, though,
we have discovered more than this. Changing the posi-
tion of the decimal point doesn’t affect the actual digits
in the sum, provided the decimal point is moved in the
same way in each number involved in the addition. If
we moved each decimal point one place to the left, we
would have 0.17 + 0.14 = 0.31; and if we moved each
one place to the right, we would have 17+ 14=81.
(This answers the question I asked on page 21 as to
how the addition rule could be used to add 17 and 14.)
Suppose we want to add 83 and 56. We have already
determined (Figure 17) that 8.8 + 5.6= 19.9. We can
tell at once, therefore, that 83 + 56 = 139.
You see, then, that the addition rule, subdivided in
decimal fashion, can be ug to add, with one shift of
boon Figure te
the slide, any two-digit number to any other two-digit
number to give sums up to 200. This is better than merely
reaching 20, but 200 isn’t thé highest number there is,
either.
Suppose we divide each tenth on the scale into ten
still smaller subdivisions, We would have tenths of
tenths, or hundredths. We could then mark off numbers
like 5.23 and 1.81 and we could find that 5.23 + 1.81 =
7.04. By shifting decimal points mentally we would find
that 52.3 + 18.1 = 70.4 and that 523 + 181 = 704. We
could now add any three-digit number to any other three-
digit number and get sums up to 2000.
But there’s a catch, and that is that there's a limit to
how finely you can divide your units. If you try to divide
the space between the units into hundreds, then you will
haye tiny markings spaced a hundredth of an inch apart,
and these would be far too difficult to read without close
peering.
‘Actually, the smallest divisions into which an inch-
a Figure 17handle, Suppose, for instance you wanted to add 3%
and 5%.
‘The decimal equivalent of 8% is 8.125, and you have
to estimate the position of 8.125 as an imaginary marking
just one quarter of the way from 3.12 to 3.14.
As for 5%, that, in decimals, is 5.8833333 . . . and so on
for any number of 3's you care to write. How do you
locate that on the addition rule? You have the markings
for 5.82 and 5.34, and the imaginary line halfway be-
tween is 538, Well, one-third of the way from 5.38 to
5.84 is the number you want.
Very well, then, you place S-0 on the B-3.125 and look
under the $-5.383888 . . . and find yourself at a position
just short of B-8.46, The position is somewhere between
that mark and the imaginary mark that represents 8.45.
You might estimate that it is three-quarters of the way
over from 8.45 to 8.46 and therefore decide that the
answer ought to be, well, 8.458.
But you have been estimating in three places now.
You've estimated for 3.125, and for 5.33333, and now for
8.458. What is the real answer? If you work out the sum
of 3% and 5% with pen and paper, you will find that the
answer is 8'%+ and if you change that into decimals it
comes out to 8.458333333 . . .
‘The addition rule did not give you the right answer.
Look how close it came, though. It was off by only
0.0008333. You may feel hardhearted and say, “A miss
is as good as a mile and a wrong answer is a wrong
answer.” But is it?
In the first place, in using the addition rule, you could
find the almost-right answer in a moment, in a fraction
of the time it would take you to get the exact answer by
pen and paper. a
To be sure, if you simply must have the exact answer
and nothing else, the addition rule will have failed you,
but often you can make do with less. It frequently
happens in science, engineering, architecture, or in any
field where numerical calculations are much used, that
it is not necessary to get the absolutely exact answer. A
very close answer will he fine. In that case, the addition
rule with its close answer is what you need.
Then, too, even if you want the exact answer, the
addition rule can still be useful as a check. Suppose you
added 3% and 514, got the correct and exact answer of
and wanted to change that into decimals. You
might perhaps get the answer 8.43842 through some
arithmetical mistake involved in long division. (It is
easy to make arithmetical mistakes in long division.)
If you then perform the same addition on the addi-
tion rule and find the answer is “just short of 8.46,” you
know something is wrong. If the correct answer were
8.43842, then the addition rule would say the answer
was “just short of 8.44.” The addition rule, if properly
handled, cannot mistake 8.46 for 8.44, and you know
you have made an arithmetical error in your pen-and-
paper calculations and start checking it. So you see, the
addition rule can be extremely useful even when it gives
only approximate answers and not exact ones.
And, of course, it is important to realize that all through
this book, I will be giving approximate answers. When
I say that 3% + 5% = 8.458, it is not intended to be an
exact answer but an “addition-rule answer.” It is suffi-
cient that it is a very close answer.
Reversing the Process
And what about subtraction? Since, in arithmetic,subtraction is the opposite of addition, it would seem that
any device that is capable of working out additions can
work out subtractions, too, if it is run backward.
On page L1,for instance, I pointed out that you could
solve 10-8 on an ordinary ruler by counting off 10
inches and then moving back 8 inches to the figure 7.
Can't that be done on the addition slide rule?
Your first impulse might be to place S-0 over B-10,
‘but what good would that do us? We can’t count back-
ward (that is, to the left) on the slide under such con-
ditions since there is nothing to the left of S-0.
One way out would be to move some other part of
the slide over the B-10. We might move S-7 over it and
count back to $-4, or move S-5 over it and count back
to S-2, Best of all, we might move $-3 over it and count
hack to S-0.
You can see that the last alternative is best, because it
is autgmatic. The 3 is part of the problem, 10 — 3, and
you are, in effect, subtracting the 8 from itself, leaving 0.
Whenever you subtract a number from itself, you leave 0,
ratio >
so that if you follow this procedure, you will always find
the answer under $-0.
This fits in neatly with the process of addition, if we
remember that subtraction is the reverse of addition.
Consider two problems: 5 + 2 and 5 — 2. In the former,
we place $-0 over B-5 and look forward to $-2, under
which we find B-7. In the latter, we place $-2 over B-5
and look backward to $-0, under which we find B-3
(Figure 20) Thus we find that 5 + 2 = 7and5—2=3.
Here we run the danger of falling into a difficulty. In
addition, it doesn’t matter in which order we add our
numbers. Faced with the problem of 5 + 2, it doesn’t
matter whether we treat it as 5+2 or as 2+5. The
answer will be 7 in either case. This lends the addition
rule a certain flexibility, If you place $-0 over B-5, the
answer will appear under $-2, If you place $-0 over B-2,
the answer will appear under $-5. In either case, you will
find B-7 as the answer.
In subtraction, matters are less flexible. The answer to
5 —2 is not the same as the answer to 2—5, In theformer case, the answer is 3, in the latter it is —3.° We
must, therefore, be careful about the manner in which we
manipulate the addition rule if we are to solve subtraction
problems correctly.
‘The best system 1 know of is to remember that the B-
numbers are immovable while the S-numbers can be
pushed back and forth and can therefore be viewed as
changeable. In the problem 5— 2, the first number, 5
(the minuend), is the number you start with; a number
which, if untouched, will remain 5 forever. It, therefore,
belongs among the unchanging B-numbers. The second
number, 2 (the subtrahend), introduces change; its
presence alters the value of 5. The subtrahend, therefore,
* It fs possible to design an addition rule to deal with negative
numbers. To do this we would have the B numbers and $ num-
bers extend leftward from 0 to give a series of ten negative
numbers. You can then have readings of B-(~5) or S-(~3),
for instance, However, such # negative extension will have no
application to the real slide rule and its uses, which 1 will be
dealing with shortly. We will, therefore, psy no attention to
negative numbers.
Figure 21
belongs among the changeable S-numbers.
In the general subtraction problem, a—b =c, then,
you place $-b over B-a and then find B-c under S-0.
Following this system, you will see that to solve 5.84 —
2.28, you place $-2.28 over B-5.84 (never vice versa),
and under $-0, you will find B-3.56. You know, therefore,
that 5.84 ~ 2.28 = 3.56 or that (just as surely) 584 — 228
356.
‘There's another way of handling subtraction. Suppose
the numbers on the slide ran in reverse, with the 0 on the
extreme right, and then read leftward, 1, 2, 8, and so on,
all the way to the 10 on the extreme left. This would
be an inverted scale, in place of the ordinary direct scale.
If you place the 0 of such an inverted scale over a B-
number, you could then find room to look leftward and
perform a subtraction.
We can call such a range of inverted numbers on the
slide, SI-numbers (for “slide, inverted”). If we wantedFigure 22
to solve 5 —2 using SI-numbers, we would place SI-0
over B-5 and, looking under SI-2, find B-3. If we wanted
to add with the SI-numbers, we could solve 5 + 2 by
placing SI-2 over B-5 and finding B-7 under SI-O (Figure
24)
Tf you compare Figures 20 and 21, you will see that
whether you use S-numbers or SI-numbers, you move to
the right (forward) in addition and to the left (back-
ward) in subtraction. The difference lies in the use of the
0. In addition, you place the S-0 over the number to
which you are adding, but not SI-0. In subtraction, you
place SI-0 over the number from which you are subtract-
ing, but not $-0. If, then, you want to make use of 0 in
this fashion for both addition and subtraction, you must
use the S-numbers for the first and the SI-numbers for the
second.
‘This does not mean we must have an addition rule and
a subtraction rule, as two separate devices. There is room
for both sets of numbers, the § and the SI, on the addition
rule we have been using
After all, there are two parts to the body of the addition
rule, an upper part and a lower part, and so far we have
been using only the lower part. There’s nothing to pre-
vent us from using the upper part, too.
Let’s place a set of numbers on the upper part of the
body, exactly like that on the lower part of the body.
‘We can refer to the upper set as the UB-numbers (“upper
body”) to distinguish it from the ordinary B-numbers we
have been using till now.
Along the upper edge of the slide, adjacent to the UB-
Figure 23numbers, are the SI-numbers. Thus we have two pairs of
scales on the same addition rule, one for addition with S-
numbers and one for subtraction with SL-numbers (Fig-
ure 22.)
Let's use our double addition rule to work out 6 +8
and 6 — 3 just to see how it works,
For addition, place S-0 over B-6 and look under S-3,
where you will find B-9, so that you see that 6+ 3=9.
For subtraction, place $I-0 under UB-6 and look over SI-
3, where you will find UB-3, so that 6— 3 =8 (Figure
23).
This brings to an end what I have to say about the
addition rule, There is considerably more that could be
written, but I now have all I need for the main business
of this book.
‘As I said in the first chapter, addition rules are not
manufactured, because addition and subtraction are such
easy operations with pen and paper (except for occa-
sional fractions) that no one bothers to take the trouble
to manipulate an addition rule instead.
Please don’t feel cheated at this. Don’t feel you have
learned various manipulations “for nothing.” All the
principles I have described in connection with the addi-
tion rule can also be used on actual slide rules which
deal with operations more complicated than addition and
subtraction.
On actual slide rules, however, you might have had
trouble seeing why those principles worked, because the
scales used are tricky ones. As it is, you learned the
principles on the simplest possible scales, those made up
of evenly spaced digits. I hope that it will now be easy
to transfer the principles from the simple scales of the
addi rule to the complex i ssiles of the true slide rules,
3
Logarithms
Multiples of Two
Acrvat.x, the whole key to any slide rule, however
fancy and complicated it may seem, is only addition and
subtraction. The point is, however, that you must add
and subtract special kinds of numbers. In this chapter,
we will track down those special kinds
Let us look, for instance, at the following set of num-
bers: 2, 4, 8, 16, 32, 64, 128, 256, 512 You can con-
tinue the set as far as you wish, for, as you see, each num-
ber is just double the one before.
‘An odd thing about such a list of numbers is this if
you multiply any two of them, the product will be
another number on the list. Thus, 2 X 8 = 16; 16 x 82 =
512; 4 X 64= 256. Then, too, if we want to get into
larger numbers, 512 X 512 = 262,144; and if you con-
tinue to work out the list of numbers by doubling each
new value, you will find that the eighteenth number on
the list is indeed 262,144.
This list (or “set”) of numbers is, in other words,
“closed to multiplication.”
Let's look at the list in another way. Since we get each
number by doubling the one before — that is, by multi-
plying the previous sims by. two —we can write thefirst number as 2, the second as 2% 2, the third as
2X 2X 2 the fourth as 2X 2X 2X2, and so on. The
list becomes a set of multiples of 2.
Of course, it is clumsy to list the numbers as more and
more 2's multiplied together, and it is natural to search
for a simpler means of indicating this. The system com-
monly used is to write 2, standing by itself, as 2°. The
number 4, which is 2X 2 (two 2's multiplied together),
is 22, The number 8, which is 2 X 2 X 2 (three 2's multi-_
plied together), is 2, By this system, you would expect
2 tobe 2X 2X 2X 2X 2 (five 2's multiplied together )
and if you work out the product you will find it to be 82.
Therefore, 2° = 32.
‘The small * (in the expression 2°) is an exponent and
25 is an exponential number. Suppose, then, we try to
‘express our set of numbers in exponential form. Instead
of writing 2, 4, 8, 16, 32... we can write 2', 24, 2
The latter method is clearly the simpler and neater of
the two.
Suppgge we multiply these numbers in their expo-
nentiff forms. Instead of writing 28 = 16, we can
waflite 2! X 2* = 2, Instead of 16 X 32 = 512, we have
2" X 9 = 2. Instead of 512 X 512 = 262,144, we have
2” x 2"
If you will examine these multiplications, you will see
that, in every case, the exponents have been added. In
the case of 2? X 9*= 24, 1+8=4. In the other two
cases 4+5=9 and 9+9=18.
This is not mysterious. In the case of 2* X 2°, you are
multiplying a set of four 2’s multiplied together (24 =
2X 2X 2X2) by a set of five 2's multiplied together
(2 =2X2X2X2xX 2). If you multiply the first set
by the second set, you end ath a set of nine 2's multi-
plied together:
(2X 2X 2X2) X (QKWKWKW’KI)=
(2X2K2XQ2KBWKIK WK IK)
In multiplying sets of 2's in this fashion, you add the
number of 2's in the various sets, and this is why expo-
nents are added when exponential numbers are multi-
plied. We can inake this general by saying that 2" X 2" =
get”,
Next, we'll try something else. Take a number of our
set of multiples of two and divide it by a smaller number
of the set. The quotient is also a number of the set. Thus
256 + 32 = 8, Turn that into exponential numbers and it
is 2° + 2" = 2", As you might guess, 8 —5 =3.
This also is not surprising. If we divide 2° by 2°, we are
carrying through the following division:
2XQXWKAXAX*KXIWXF
2X2XQXQXD
ite
The five 2's in the denominator cancel five of the eight =.
2s in the numerator, leaving three 2’s multiplied together
as the answer. Therefore, 2* + 2° = 2* and, in general
ge grage-ne
We have thus discovered a method of converting cer-
tain multiplications into additions, and certain divisions
into subtractions. Since it is the exponents which are
added and subtracted in these cases, we are going to
concentrate on these exponents from now on.
If we extract the exponent from an exponential number
and set it down all by itself, it is customary to give ita
new name and call ta logarithm. For instance, if we areconsidering the expression 2" = 8, then 3, taken by itself,
is the logarithm of 8. Of course, we don’t want to forg:
the 2, which rests under the exponent as though it wer
a base. Consequently we speak of 8 as the logarithm of
8 to the base 2.
In the same way, the logarithm of 16 to the base 2 i:
4, for 16 = 2'; and the logarithm of 82 to the base 2 is 5,
for 2* = 82.
‘We can abbreviate “logarithm to the base 2” as logs. Ir
that case, we can say: log: 8 = 8, logz 16 = 4, log: 32 = 5,
and so on.
Suppose we want to work it backward now. We hav
the logarithm to the base 2, and we want to write down]
the number it represents, If we have the logarithm 4 to}
to the base 2, then the number it represents is 16, so that}
16 is the antilogarithm of 4. In the same way if we have
the logarithm 9 to the base 2 then its antilogarithm is 512,
sinc? 2° = 512, The antilogarithm in these cases is to the|
base 2, andthe expression can he abbreviated antilogs.
Ths, we can say that antilog: 4 = 16, and antiloge 9=
“Sr.
These new terms are confusing at first and will cease’
to be confusing only with practice. However, to be as
clear as possible right now at the start, consider the ex-
pression x* = b. In this general exponential expression, x
is the base, a is the logarithm, and b is the antilogarithm.
(It might help you to keep in mind the fact that log-
arithms are exponents, while antilogarithms are the “ordi-
“*"This relationship between multiplication and division of ex-
‘ponential numbers — where exponents are added in the first case
‘and subteacted in the second — is not surprising. Division és an
‘operation which is the inverse of multiplication, just as subtrac-
tan is the inverse of addition. Keep this in mind for it will be
Jrandy later.
hary numbers” you use in everyday computations. )
Jt is easy to make a small table of logarithms and anti-
Jogarithms to the base 2, as follows:
loge antilog: logs antilog:
1 2 ul 2,048
2 4 12 4,096
3 8 13, 8,192
4 16 4 16,384
5 82 16 82,768
6 64 16 65,536
7 128 iT 181,072
8 256 18 262,144
9 512 19 524,288
10 1,024 20 1,048,576
You can continue the list as long as you like, but we
have enough now to make our point. Remember that
whenever we multiply two antilogarithms, we @gr achieve
the same result by adding the corresponding loners
Suppose, for instance, we wanted to multiply 128 aff",
4096, These “ordinary numbers” are antilogarithms and
may be found in the column headed antiloge. The loga-
rithms corresponding to them are 7 and 12 respectively.
We add the logarithms and find that 7 + 12 = 19, so that
19 is our logarithm sum. We find 19 in the logz column,
and sce that 524,288 is the corresponding antilogarithm.
‘The logarithm sum is the antilogarithm product. We
therefore conclude that 128 4096 = 524,288, a fact you
can check by long multiplication.
‘Again, since 10 + 8 = 16 (logarithms), we can see at
once that 1024 X 256 = 262,144 (antilogarithms. )
We can try subtracting logarithms, too. Since 20 — 15(logarithms), we conclude that 1,048,576 + 32,7
82. Try this by long division and see if it is not righ
You will agree, I think, that it is much easier to add a1
subtract logarithms and make use of columns such
those given above, than it is to multiply and divide or}
dinary numbers (that is, antilogarithms ).
Multiples of Ten
There are, however, flaws to this pretty picture. Let's
begin with the biggest flaw of all. What I have descril
so far, will only work for a few numbers which happen
be built up through the multiplication of 2's, You ca
deal with 32 X 64, but suppose you want to deal wit
81 X 63. You are stuck. You can’t produce either 81
63 by multiplying 2's.
‘One possible help might be to produce numbers whic
are multiples of other integers. For instance, you can't
produce either 27 or 81 by multiplying 2's, but you cz
produce them by multiplying 3's. Thus, 27=3 X33
8° and 81=3X8 8X 8=38'. If you want the answe
to 27 X 81, try it in exponential form. Since 3 + 4= 7,
3° +8'=8". The number 3" represents the product of
seven 3's multiplied together, and that works out to}
2187. So you can say that 27 x 81 = 2187,
In short, you can prepare columns that list logarithms
to the base 8 and their corresponding antilogarithms t
the base 8. You can then solve problems which you can’t
solve by using logarithms to the base 2 in the fashion
described above.
In the same way, we can work with logarithms to d
base 5, or to the base 7, or to any base we may care
choose and in each case work with a new set of numbers
4
However, this does not solve our problem. Every
number can be found among the list of antilogarithms to
‘one base or another, but what if you want to multiply a
number from one list by a number from another list?
Consider, for instance, the problem 8X 9=72. You
can see that 8 = 2" and 9=8*. This means that log:
8 = 8, and logs 9 = 2. These are logarithms to different
bases. Does that matter? If they were logarithms to the
same base, we would add them and achieve the same
result that we would by multiplying the antilogarithms
8and9. Let's try adding them anyway. Since 2+3=5,
we decide that 5 is the logarithm of the product — but
the logarithm to which base?
If 5 is a logarithm to the base 2, then the answer is 2°
or 82. If it is a logarithm to the base 8, the answer is 3'
or 243, But the actual answer, 72, is neither. In short, we
simply can’t add logarithms to one base and logarithms
to another base, any more than we can add boys to cows.
If we are going to make logarithms useful, we must
stick to one particular base and work out a way for
finding logarithms to that base for all possible numbers.
But if we are to do this, which particular base are we
to use?
Perhaps at first, you might think that the most con-
venient possible base is 2, and in some ways it is, After
all, doublingis so simple that it is much easier to prepare
a list of multiples of 2 (2, 4, 8, 16, 32. . .), than of 3's
(3, 9, 27, 81, 243 . ..), or of 5's (5, 25, 125, 625, 3125...)
or of almost any other number. Furthermore, the num-
bers in the list of multiples of 2 are more closely spaced
together than the numbers in the list of multiples of any
other numbers, so there are fewer numbers not on the
45We must therefore extend the notion of logarithms t
numbers that don’t end with zeros.
Let's begin by considering the following: 10 + 1
We can express 10 in exponential form as 10" and thi
express the division as follows (remembering to subtra
exponents): 10'+ 10'= 10°. If we solve a problem,
correctly, in two different fashions and get an answer in
two different forms, those forms ought to represent the
same number. Here we divide 10 by 10 and get
answers: 1 and 10°. It is reasonable then to suppose th:
10°= 1. If we bring the exponent down, it bec
a logarithm and we can say that log 1= 0.
We know now that log 1 =0 and log 10= 1. What
about the logarithms of numbers between 1 and 10? I
would seem that those logarithms ought to lie between|
0 and 1 in value. They ought, in other words, to be|
fractions.
But what meaning could a fractional logarithm have?|
Suppose there were a number which had a logarithm
equal to %. What kind of number would fit such
logarithm?
Let's consider a number x, such that log x = Re-|
member that a logarithm is simply an exponent brought]
down from its position above. If log x = '4, then x = 10",
We are certainly entitled to wonder what a number like]
10% can possibly mean, Suppose we consider the produc
of 10% X10". If we add exponents we can see that
10% x 10% = 10' = 10.
We can therefore say that 10% is that number whit
when multiplied by itself, gives 10 for an answer. Sucl
anumber is the “square root of 10” and is usually writte
‘10.
It is not difficult to work out an approximate value
8
the square root of 10. Actually, it tums out to be an un-
ending decimal expression, but it is possible to work it
‘out to as many decimal places as are desired, and the
more decimal places that are worked out, the closer the
number is to the actual value of the square root of 10.
Worked out to six decimal places, the square root of
10 is 3.162120, It turns out that 8.162120 x 3.162120 =
9.9990028044, which is almost 10 as you see. In fact, it
would be pretty useful to speak of the square root of
10 as 3.162, This number, multiplied by itself, yields the
product 9.998244, which is still pretty close to 10.
Since 3.162 X 3.162 is just about equal to 10, we can
say that 8.162 is just about equal to 10, We can there-
fore say, with a good approximation to the truth, that
antilog '4 = 8.162 and that log 3.162 = %.
Next let's consider that 10 x 10% x 10% = 10' = 10.
Therefore, 10% is a number which yields 10 when mul-
tiplied by itself @Ase/4We can, therefore, say that 10%
is equal to the “cube root of 10” or 10. It is possible
to work out the cube root of 10 and this turns out to be
(to three decimal places) 2.154. Therefore we can say,
as an approximation, that antilog % = 2.154 and that
log 2.154= %.
By adding exponents, we can see that 10% x 10¥ =
10™. Wealready know that 10* is equal to approximately
2.154, If we multiply 2.154 by itself we get about 4.64.
Therefore, 10% is equal to about 4.64, which means we
can say that (approximately, at least) antilog % = 4.64
and log 464 = %.
You can, if you choose, express such fractional loga-
rithms in decimal form; and, indeed, they usually are so
expressed. Instead of saying log 3.162 = %, you would
49list for which we must find logarithms.
However, here as in many other cases, the fact that
our number system is built upon the number 10 over-
rides everything else. Because our number system
built upon 16, it is easier to prepare a list of multiples o!
10 than of any other number.
We begin with 10 itself. Then we have 10 X 10 = 100.
After that is 10 X 10 X 10 = 1000; 10 X 10 X 10 x 10
10,000, and so on, Our list is 10, 100, 1000, 10,000 . .
each number possessing one more zero than the numl
before. There is nothing easier than adding one
zero for each number on a list.
‘We can express such a list in exponential form, too,
10°, 10%, 10*, 10°... The connection between the ordi-
nary 10-multiple and its exponential form is a simple one_|
Since 10 = 10', 100 = 10 X 10 =102, 1000 = 10 x 10 x}
10 = 10", and so on, you can see that the exponent i
equal to the number of zeros in the number itself. You
can see without having to go into great feats of computa-
tion that 10,000,000 = 10° and 100,000,000,000 = 10".
The simplicity of this connection extends to logarithms
to the base 10, Remember that the logarithm is the ex-
ponent. Therefore since 100 = 10%, logio 100 = 2. Again,
since 10,000,000 = 107, Logo 10,000,000 = 7. Again, the}
logarithm to the base 10 is equal to the number of zeros|
in a number of this sort.
Such logarithms are so easy to work out that loga-
rithms to the base 10 are used far more often than loga-
rithms of any other kind.” Consequently, when people
"There is a set of logarithms to another base very frequently,
used i higher mathematics, Since this book does not involve
higher mathematics, we will not need to consider this other set,
semetimes tere oto lgatins”
speak of “logarithms” without specifying the base, they
are almost certain to mean logarithms to the base 10.
For the rest of the book, logarithms to the base 10 are
the only ones I shall use and I shall refer to them merely
as logarithms, with the simple abbreviation log.
‘Logarithms of this sort are so easily obtained that one
doesn’t even need a set of columns of logarithms and
antilogarithms to carry through multiplication and
sion of multiples of ten. Suppose we consider 10,000 X
100. By counting zeros, we know at once that log
10,000 = 4 while log 100 = 2. Since 4 + 2= 6, we know
the logarithm of the product to be 6. That means there
are 6 zeros in the corresponding antilogarithm, so that
10,000 X 100 = 1,000,000.
‘Again, if we consider 10,000 ~ 100, we must subtract
logarithms. Since 4 — 2 = 2, the logarithm of the quo-
tient is 2 and the corresponding antilogarithm is 100.
Therefore 10,000 + 100 = 100.
Between the Multiples
Of course, we can't be completely enthusiastic about
all this, for so far we can only make use of the 10-
multiples, such as 100 and 1000, and these are far fewer
than the 2-nultiples. If you inspect the columns on
page 45, you will see that there are twenty 2-multiples
up to the neighborhood of a million. There are only six
10-multiples in that same range.
Furthermore, the 10-multiples are particularly easy to
handle even without logarithms. It is no problem at all
to decide, without logarithms, that 100 X 1000 = 100,000.
What we need is some easy method of carrying through
a multiplication such as 72 FS 263.say log 3.162 = 0.5. If you worked out the fact that log
1.585 = %, you could express it as log 1.585 = 0.2, and
soon.
The cases I've presented above are examples of ways
in which simple fractional logarithms can be obtained.
Those simple fractions have as their antilogarithms
rather complicated decimals. Mathematicians, however,
have worked out methods for obtaining the logarithm
of any number at all, including logarithms for all the
simple digits. Thus, the logarithm of 2 turns out to be
about 0.801 (to three decimal places), so we can say
log 2-= 0.301.
Actually, the logarithm of a number, in almost every
case, is an unending decimal, but mathematicians can
work out as many places as they choose. They can then
prepare logarithm tables, in which the logarithms of a
set of consecutive numbers are presented. Such loga-
rithms are commonly presented to five decimal places.”
By using such a table, we can find that:
antilog log
1 0.00000
Secwmrsaunen
2
at
e
a
This is equivalent to saying that we can place each
50
of the digits from 1 to 10 into exponential form. We can
say that = 1047712, 7 = 10°**97° and so on.
The usual five-place logarithm table will, of course,
also give the logarithms of decimal numbers. For in-
stance, such a table will tell you that log 4.354 = 0.63889
and that log 2.189 = 0.84025. That means that 4.354 =
10°**#*° and that 2.189 = 10°02,
In multiplying exponential numbers, we add expo-
nents — which is to say, we add logarithms. Suppose, for
instance, you wanted the answer to 4.354 X 2.189. In-
stead of multiplying the numbers themselves, you add the
logarithms and find that 0.63889 + 0.4025 = 0.97914.
The sum 0.97914 is the logarithm of the product of 4.354
X 2.189, In the logarithm table, we find that 0.97914 is
the logarithm of 9.531. Therefore, we conclude that
4.354 X 2.189 = 9.531.
This is not the exact answer. The manipulation of
logarithms can’t give you the exact answer unless the
exact values of the logarithms are used, and logarithris
are virtually never known exactly. However, five-place
logarithms are accurate enough for most purposes. For
instance, if you work out 4.354 X 2.189 with pencil and
paper, you find the exact answer is 9.530906 and surely
9.531 is quite close,
In the same way, you can subtract logarithms instead
of dividing numbers. Suppose you wanted the answer to
4.354 = 2.189, Taking the logarithms again, we find that
0.63889 — 0.84025 = 0.29864. The logatithm-difference
0.29864 is the logarithm of the quotient of 4.354 + 2.189.
“16 this were @ book on logarithms, 1 would include euch &
table and give complete instructions for its use. However, Tam
txing logarithms only to explain the workings of the slide rule.
Twill ask you, therefore, to accept, more or less on faith, the
values of the logarithms as T present them in the meanwhile,
a‘The logarithm table tells us that 0.29864 is the logarithm
of 1.989. Therefore, we conclude that 4.854 + 2.189 =
1.989. If you work out the problem by long division, you
find the answer is actually about 1.98858, but again 1.989
is reasonably close,
With practice, one could learn to use logarithm tables
so quickly and easily that there would be no question of
working out complicated multiplications and divisions in
the ordinary way—unless one simply had to have
absolutely exact answers. Even then, the logarithmic
answer would be a convenient check against the ever-
present passibility of arithmetical error.
Changing the Characteristic
So far, we have been working with the logarithms of
the numbers from 1 to 10. There is, however, a vast
array of numbers greater than 10 and smaller than 1.
What about those?
What might the logarithm of 52.88 be, for instance?
We can solve that problem by considering 52.38 as
5.238 X 10. The logarithm of 5.288 is, according to the
logarithm table, 0.71917. The logarithm of 10 is, of
course, 1. If we are multiplying two numbers to get a
product, we can just as well add the logarithms of those
two numbers to get the logarithm of the product. Since
5.288 X 10 = 52.88, and 0.71917 +1= 1.71917, then
1.71917 is the logarithm of 52.88. We can express this
more simply by saying log 52.88 = 1.71917.
It is simple to find the logarithm of 523.8 now. We see
at once that 5.238 X 100 = 523.8, and we know that the
logarithm of 100s 2. Therefore using logarithms, 0.71917
52
+2=2.71917 and log 523.8 = 2.71917. In the same
way, we can show without trouble that log 5238 =
3.71917, that log 52,380 = 4.71917, and so on.
Let's make a small table to make sure we sce the point
clearly:
antilog log
5.238 0.71917
52.88 L71917
523.8 2.71917
5,238 8.71917
52,380 4.71917
We can divide the logarithm into two parts: the num-
ber to the left of the decimal point, which is called the
characteristic, and the number to the right of the decimal
point, which is called the mantissa. If we say that log
52.38 = 1.71917, the characteristic of that logarithm is 1,
the mantissa is 71917.
When two numbers differ only in the position of the
decimal point, the mantissa of their logarithms is iden-
tical, as you see in the example given in the table above.
The characteristic, on the other hand, changes with the
shifting decimal point in a number.
‘The change in the characteristic presents no problem.
By inspecting the table above you can see that in the
case of those numbers, at least, the value of the character-
istic is one less than the number of digits to the left of the
decimal point in the antilogarithm; and this it turns out is
truc in all cases.
This rule is so simple that we need no logarithm table
to determine characteristics, By simply looking at the
number 35.62, we see that there are two-digits to the left
53of the decimal point and that the characteristic is there-
fore 1. In the same way we know that 282,100,000, which
has nine digits to the left of the decimal point, has a
characteristic of 8.
All we need the logarithm table for, then, is to de-
termine the mantissa, and it should, therefore, really be
called a “mantissa table.” Since the mantissa is the same
for any particular combination of digits, regardless of
the position of the decimal point, we can ignore the
decimal point altogether in finding the mantissa.
Suppose we want the logarithm of 35.62. We look up
the mantissa for the digit-combination 3562 and find it to
be 55169. Since 35.62 has two digits to the left of the
decimal point, the characteristic of its logarithm is 1; and
therefore log 35.62 = 1.55169.
In the same way, to get the logarithm of 282,100,000,
we look up the mantissa for the digit-combination 2821
and find it to be 45040. Since 282,100,000 has nine digits
to the left of the decimal point, the characteristic of its
logarithm is 8. Therefore, log 282,100,000 = 8.45040.
Suppose, then, we wanted to multiply 123.1 and 35.2.
We know at a glance that the characteristic of the loga-
rithms of these two numbers are 2 and 1 respectively.
We find the mantissas of the two digit-combinations 1231
and 3520 in the log tables and find them to be, re-
spectively, 09026 and 54654. Therefore log 123.1 =
2.09026 and log 35.2 = 1.54654. If we add these two
logarithms, 2.09026 + 1.54654, we get 3.63680, which
is the logarithm of the product of 123.1 X 35.2.
The problem, now, is to find the antilogarithm of
8.63680. It is the mantissa that will give us the necessary
digit-combination, and the logarithm table tells us that
the digit-combination of the antilogarithm represented by
4
the mantissa 63680 is 4333. The characteristic, 3, tells us
that we must place the decimal point so that there are
four digits to its left. The answer is, therefore, 4333.
(The actual answer, obtained by multiplying 123.1 and
35,2 in full, is 4833.12. )
We can divide similarly. If we want to divide 123.1 by
35.2, we subtract the logarithm of the latter from that of
the former. It tums out that 2.09026 — 1.54654 =
0.54372. The digit-combination of the antilogarithm of
the mantissa, 54372, is $497. Since the characteristic is 0,
the decimal point must be so placed as to have one digit
to its left. We can therefore say that 123.1 + 35.2=
3.497. (If we work out 123.1 + 35.2 by long division, the
‘answer turns out to be about 3.49715.)
‘And what about numbers less than 1? Suppose, for
instance, we are faced with determining the logarithm of
0.481. We use the same strategy that worked for us be-
fore. We will consider 0.481 as 4.81 + 10. The logarithm
of 4.81 is 0.68215 and the logarithm of 10is 1. In dividing
numbers we must subtract logarithms to get the logarithm
of the quotient. Therefore, if 4.81 + 10=0.481 and
0.68215 — | = —0.31785, then —0.31785 is the logarithm
of 0.481.
This presents us with a slight problem. The mantissa
of —0.31785 is 31785, which is different from the mantissa.
of the logarithm of 4.81. Furthermore the characteristic
is 0, whatever that means, Fewer problems are created
if we leave the logarithm of 0.481 as 0.68215 — 1, without
trying to carry through the subtraction. That leaves the
mantissa the same as in the logarithm of 4.81 or 48.1 or
any of that family, And we can call —1 the characteristic.
In the same way, we find that 0.0481 can be written as
481+ 100, so that its logarithm is 0.68215 — 2, the
55logarithm of 0.00481 is 0.68215 — 8, and so on.
If we write positive numbers less than 1, with one zero
to the left of the decimal point, we can then work out a
simple rule for determining the characteristic of their
logarithms. The characteristic is a negative number with
its value equal to the number of consecutive zeros that
begin the antilogarithm.
For instance, 0.235 begins with a single zero, and
the characteristic of its togarithm is —1. The number
0.000442 begins with four consecutive zeros and the
characteristic of its logarithm is —4.
It is only consecutive zeroes that count in this rule. The
number 0.02004, has four zeros all together, but it starts
with two consecutive zeros and its characteristic is —2.
We find the mantissa in the ordinary way. For 0.235,
we look up the mantissa of the digit combination 235 and
find it to be 7107. Since the characteristic is —1, log
0.285 = 0.87107 — 1.
Again, to find the logarithm of 0.000442, we look up
the mantissa of the digit-combination 442, This turns out
to be 64542, so that log 0.000442 = 0.64542 — 4, To get
the logarithm of 0.02004, we look up the mantissa of 2004
and find it to be 30190, so that log 0.02004 = 0.30190 — 2.
We can make use of these logarithms with negative
characteristics in quite the usual way. If we wish to
multiply 235 by 0.000442, we need but add the log-
arithms: 2.87107 + (0.64542 — 4) and this comes out to
3.01649 — 4, or, simplifying it, 0.01649 — 1. The man-
tissa 01649 corresponds to the antilogarithm digit-
combination 10887 and from the characteristic of —1,
we know the decimal point should be placed so as to
make the answer 0.10387.
If we were considering the problem 235 + 0.000442,
we would subtract the second logarithm from the first.
This means 2.87107 — (0.64542 — 4). As you probably
know, subtracting a negative number is the same as
adding a positive one, so that by removing parentheses,
we have: 2.37107 — 0.64542 + 4 = 5.72565. The man-
tissa 72565 gives us 5317 as an antilogarithmic digit-
combination, From the characteristic 5, we know that the
answer must have six digits to the left of the decimal
place and must therefore be 581,700.
Let us summarize, then. For any positive number, we
can work out the characteristic of the logarithm at a
glance. The mantissa can be found in a logarithm table
and will be the same for any particular digit-combination
regardless of the position of the decimal point.”
“This is tre, in our number system, only for logorithms to the
base 10 — anther reason for preferring the base 10 to any ather
in ordinary computations, Nor is this a coincidence. It is the
inevitable consequence of the fact that our number system is
Dail around 10 and maltiples of 10,
a7
—4
Logarithms on Wood
The L-Scale
Le us stop now and consider the situation. In Chapter
2,.1 showed how we could add and subtract numbers,
automatically, on an addition rule. In Chapter 3, 1
showed how we could multiply and divide numbers by
adding and subtracting their logarithms.
Could we not combine these two processes? Could we
not use an addition rule to add and subtract logarithms
and thus carry through multiplication and division auto-
matically? «
Indeed we can. An addition rule used to add and sub-
tract logarithms is the true slide rule, and it is this slide
rule which this book is intended to explain.
The simplest slide rule we can imagine is, actually,
identical with an addition rule. Suppose, for instance, we
are faced with the necessity of performing a particular
multiplication. We use a logarithm table to look up the
logarithms of the numbers being multiplied. We then
add those logarithms on the slide rule, thus obtaining the
logarithm of the product. We return to the logarithm
table to get the antilogarithm of that final logarithm, and
thus obtain the product of the multiplication.
This is clearly not very good. The time-consuming
part of working with logarithms is the finding of log-
arithms and antilogarithms in the table. Adding the
logarithms once they are found is very simple. We are
thus making the slide rule do the easy part and neglect
the tedious part.
But what if we could make the slide rule itself into a
kind of logarithm table? It might then be unnecessary to
look up logarithms and antilogarithms at all. We could
make the slide rule look them up, so to speak, and, having
found them, add or subtract them, and then give us the
antilogarithm as the answer.
Can this be done? Or is it just a beautiful dream?
Delightfully enough, it ean be done; and easily too. To
do it, we begin by marking off logarithms on the slide
rule,
For this all we need are the logarithms of the numbers
from 1 to 10, Once we have these, we can obtain the
logarithms of numbers less than 1 or more than 10 by
choosing the proper characteristic, which is easy enough
(see page 53).
The logarithm of 1 is 0.0. As we progress through
numbers greater than 1, the logarithm increases throughhigher and higher values —0.1, 0.2, 0.3, etc. —until
finally the value of 1.0 is obtained, which is, of course,
the logarithm of 10.
Suppose, then, we represent these logarithms in the
form of a line on the body of the slide rule. The line
begins at 0.0 (the logarithm of 1) and ends at 1.0 (the
logarithm of 10). This line is divided into ten equal
divisions by marks we can call primaries. The primaries
represent logarithms of 0.1, 0.2, 0.3, and so on.
In order to simplify matters and keep from blurring the
eye with unnecessary details, it is customary to omit the
zero to the left of the decimal . The primaries on this’
scale of logarithms are therefore marked off as .0, .1, .2, .3,
4, 5, 6.7, 8, 9, 1.0, from right to left. In some slide
rules the decimal point is also omitted.
This line actually deals with mantissas, rather than!
with logarithms, just as a logarithm table does. If we|
want to consider the line as dealing with mantissas to two}
places, the primaries mark off the mantissas 00, 10, 20,]
80, ete, If we wish to consider the mantissas to three|
places, the primaries are 000, 100, 200, 800, ete.; if to
four places, they are 0000, 1000, 2000, 3000, ete. 5
It will be most convenient to consider the mantissas to
three places, and we will therefore consider the primaries
as representing 000, 100, 200, 300, 400, 500, 600, 700,
800, 900, and 1000.
Between the primaries are ten equal divisions marked
off by secondaries, the middle secondary being longer
than the rest for convenience in reading mantissa values.
Between each neighboring pair of secondaries are five|
still smaller divisions marked off by tertiaries.
The result, with its numbered primaries, and its un-
numbered secondaries and tat . is shown in Figur
24, in which the scale is marked “L” (for logarithm). We
can, therefore, refer to it as the L-scale.
Notice that I have drawn the L-scale on the lower part
of the body of the slide rule in Figure 24. There is no
hard and fast rule about this. Whether the L-scale is
drawn on the upper body or elsewhere depends on the
particular slide rule.
Slide rules are manufactured in many different designs
which differ among themselves in details. The slide rule
am using as a model for the illustrations in this book has
the L-scale on the lower body. Another stide rule in my
possession has it on the upper hody. Still others have it on
the slide.
If your slide rule possesses scales in locations other
than those shown in the illustrations, do not feel dis-
turbed. The difference in location will not be important,
and you will be able, very easily, to adapt my instructions
to your sliderule.
If you have an inexpensive slide rule (and some quite
decent slide rules can be obtained for not more than a
couple of dollars) then some of the scales I will describe
& this book may be missing altogether from your instru-
ment. You will be able to follow the use of the missing
scales from the diagrams in the book, however, and with
the scales you do possess you will still be able to perform
the essentials of multiplication and division,
If you measure the length of the L-scale on almost any
slide rule, you will find that it is nearly teu inches long.
It is this which allows people to refer to a 10-inch slide
rule even though the overall length of the usual slide rule
(which naturally extends some distance past both ends of
the L-seale) is about 124% inches long,
‘The fact that the L-scale has a total length of 10 inches.means that the primaries (numbered from 1 to 10) are
about I inch apart. This makes it tempting to look upon
the slide rule as an “ordinary ruler.” This, however, is
quite wrong.
vo nor make use of the L-scale to measure lengths.
The slide rule is a precision instrument and should be
used with the greatest care and only for the purpose for
which it is designed. (Would you hitch a race horse to
a milk wagon?) Secondly, the slide rule is not adapted to
the purpose and has a blunt end against which it is hard
to measure lines. Thirdly, the L-scale is not quite 10
inches long. It is 25 centimeters long and that actually
comes to 9.84 inches, The primaries are therefore only
0.984 inches apart.
To summarize, the use of a slide rule as an ordinary
ruler is improper, inconvenient, and incorrect. Need I say
more?
It is not very difficult to read the L-scale. The reading
can be made quite easily to three places so that the pri-
mary readings, as I said before, are 000, 100, 200, 300 and
soon.
Concentrate for the moment on the primary marking’.
Figure 25 ys
3 and 4 which represent 300 and 400, The secondaries
lying between mark off tens: 310, 820, 380, 340, 850, and
soon, The long secondary in the middle is 350. This can
be told at a glance, and the shorter secondaries can be
counted off easily enough.
Between the secondaries are the tertiaries which mark
off by twos. Thus, between 330 and 340 are four tertiaries
which represent 882, 834, 386, and 888. To find 337 we
place an imaginary line midway between 336 and 338;
and we can find in similar fashion any odd three-number
mantissa.
With practice one can even estimate four place mantis-
sas. Suppose, having found 836 and 338, we consider
those two tertiaries to represent 3360 and 3880. Exactly
between lies the imaginary line marking 3370. Half way
between 3370 and 3380 would be $375 and a shade to the
right of that would be 3376 (Figure 25).
Such an estimation of a fourth place is never accurate,
however, and it is quite impossible to locate a fifth place
on the L-scale, It is for this reason that the slide rule
cannot entirely replace a logarithm table. A decent
logarithm table gives logarithms to five places at leastronfeufpanuak dant i gma
Figure 26
and thus offers an accuracy greater than that possible on
the slide rule, The slide rule makes up for its lesser ac-
curacy, however, as we shall see, by its far greater
convenience.
The D-scale
To make a logarithm table out of the slide rule, we
need more than just the mantissas of the L-scale. We
need antilogarithms, too, and we need those antilog-
arithms placed in such a way that we can in a moment
match up each logarithm with its corresponding anti-
logarithm (or vice versa).
In order to do this we need a second scale above or
below the L-scale. Each particular point on this second
scale will represent the antilogarithm corresponding to
the logarithm at the point directly above or below on the
L-scale.
To match a point on one scale with the corresponding
point on another scale directly above or below is not easy
to do by eye, however. Slide rule manufacturers have
therefore provided a mechanical device to help out.
This is an indicator assembly which consists (in most
slide rules) of two glass windows, one on either side of
the slide rule, fitted into a holder at the top and bottom
so that they are held firmly in place (Figure 26).
The indicator assembly can slide back and forth easily
along the slide rule when it is pushed. If it is left alone, it
is held in place by the friction of a spring, so that it stays
where it is put. The indicator assembly cannot be moved
off the slide rule, however, unless it is taken apart, for the
metal bar that holds the two parts of the body firmly
together‘acts to block the further movement of the indica-
tor assembly.
Down the precise middle of each glass window is a
fine vertical hairline. As the indicator assembly is moved
right and left, the hairline itself can fall over any part of
the L-scale from 0 on the far left to 10 on the far right.
The hairline can act as a convenient guide for the eye.
If you want to locate 337 on the L-scale (a reading we
can refer to as L-337), begin by adjusting the hairline so
that it falls between the tertiaries L-836 and L-338, It is
easier to put a real line between the two marks, than to
try to imagine one. By adjusting the hairline one can
even try to estimate the position of L-3376 and mark its
Position until such time as the hairline must be moved
again. The hairline, in short, finds a point on the scale
and marks it. 85It does more, too, for once the hairline is placed at
1-387, it also marks the corresponding points on various
seales above or below the L-scale. In particular, we can
find points on the scale of antilogarithms (which I am
about to discuss) that correspond to points on the L-
scale.
The indicator assembly will play a part in virtually
every drawing in the remainder of the book. In order to
simplify these drawings and in order not to obscure any
markings, I will make no attempt to picture the entire
indicator assembly. I will represent only the hairline as a
line crossing the slide rule vertically (Figure 27).
Where, now, shall we place our antilogarithm scale?
There are four key scale positions on the slide rule: on the
top and bottom of the slide, on the upper part of the
hody adjacent to the slide and on the lower part of the
body adjacent to the slide. (These are the positions
where I placed the four scales of my addition rule, See
Figure 22 on page 36.)
On early slide rules, these positions were used for the
basic scales and they were marked, from the top down,
A,B, C, and D (Figure 28). These letters are still ap-
plied to certain scales in these positions, although all
other scales are lettered by means of initials of one sort or
another, as the L-scale is lettered L for “logarithm.”
It is customary to place one scale of antilogarithms in
position D, and it therefore hecomes the D-scale, By this
arrangement, the L-scale and the D-scale are both on the
Jower body, Each maintains a fixed and permanent pési-
tion with respect to the other, since neither is disturbed
by the to-and-fro motion of the slide.”
We can begin matching the L-scale and D-scale at
the extreme left. The leftmost reading on the L-scale is
000, and this mantissa has for its antilogarithm 1, or any
number obtained from 1 by moving the decimal point —
10, 100, 0.1, and so on. Naturally, it is simplest to con-
sider the antilogarithm as 1.
If the hairline, therefore, is placed exactly on L-000, it
should intersect the left end of the D-scale and give a
reading of D-1
(If, on your own slide rule, the hairline does not
exactly intersect L-000 and D-1 simultaneously, then the
slide rule is out of adjustment. It may be that the top and
bottom halves of the body are not lined up properly, or
* Where the Lescale iso the slide, as in some rule designs, the
scale of antilogaithms can be placed in the C-position, The L-
seale and the C-scale will then have a fixed interrelationship.Figure 28
that the indicator assembly glass is not seated properly
in its metal holder. It is best not to try the adjustment by
yourself —a matter of ‘loosening screws, moving the
upper part of the body or the indicator glass, and tighten-
ing the screws again —but to allow someone familiar
with the slide rule to do so.)
Let's continue, If we refer back to the table on page
54, we will see that log 2= 0.301. (We need only use
three decimal points here, for that is all we can read ac-
curately on the L-scale.) That means antilog 0.301 = 2,
so that immediately over L-801, we place D-2. Continu- #
ing to refer to the table on page 54, we must place D-8
directly over L-477, D-4 directly over L-602, and so on.
The reading L-1000 represents a logarithm of 1.000
Figure 29
and the mantissa here is once again 000. The antilog-
arithm is, once more, 1, so that the right end of the
D-scale, like the left end, is marked 1. You see the result
in Figure 29.
Compression at the Right
T'm sure that the first thing you notice about the D-
scale is that the ten primaries marked off by the numbers
are not equal in length, The distance between 1 and 2 is
longer than the distance between 2 and 3, which is in
turn longer than the distance between 3 and 4, The
shortest distance of all, that between 9 and 10, is less than
¥, as long as the distance between 1 and 2. There is asteady compression of the distances at the right in other
words.
To see why that is, let's consider the set of expressions:
2,2X2,2X2X92,2X2X 2X 2, and soon. If we con-
sider the products of these expressions, we get the set of.
numbers: 2, 4, 8, 16, 82, ete., a set we have dealt with
before.
Next let’s consider the logarithms of the set of expres-
sions we have just mentioned. The logarithm of 2 is
0.801. If we want the logarithm of 2 x 2, we must take
the sum 0.301 + 0.801. Continuing in this way, we find
that the logarithms of the set of expressions 2, 2X 2,
2X2WX2, 2X2 2X2, ete, is 0.901, 0.801 + 0.801,
0.301 + 0.801 + 0.301, 0.301 + 0.801 + 0.501 +0.301, ete.
Taking the sums, we get the set of logarithms: 0.301,
0.602, 0.903, 1.204, and so on.
Now we can prepare a table matching the products of|
each expression in the set with the logarithm of that:
product, and we can begin with 1, which has the loga-
rithm 0.000:
antilogarithm logarithm
I 0.000
2 0301
4 0.602
8& 0.903,
16 1.204
and so on for as long as we care to continue.
Notice that the antilogarithms increase by doubling.
Each antilogarithm must be multiplied by 2 to get the
next one in the set. A set of numbers in which each is
70
obtained from the former through multiplication by a
particular number (in this case 2) is a geometric pro-
gression.
‘The logarithms, on the other hand, increase by addi-
tion. To each logarithm must be added 0.301 to get the
next one in the set. A set of numbers in which each is
obtained from the former by the addition of a particular
number (in this case 0.301) is an arithmetic progression.
It should be quite clear that the numbers in a geometric
progression increase in value much more rapidly than do
those in an arithmetic progression. You can see it’s so
in this case.
As we pass along the two sets of numbers in the table
above, you can see that a fixed change in the logarithm
brings about a greater and greater change in the anti-
logarithm, To begin with an increase of 0.301 brings the
antilogarithm from 1 to 2; another such change of 0.301
brings the antilogarithm from 2 to 4; then from 4 to 8;
then from § to 16, and so on.
A broader and broader stretch of antilogarithms must
fit into each fixed change of logarithm, which is why the
numbers on the D-scale are closer and closer together
as they get larger and larger. It is why there is com-
pression on the right.
We can show this graphically, too, as in Figure 30,
In this graph, we have a scale of logarithms from 0.0
to 1.2 on the hottom, and a scale of antilogarithms from
© to 16 on the side. The dots mark the intersections of
the logarithm vatue and its corresponding antilogarithm
value, as given in the little table on page 70. A smooth
curve is drawn through these dots and this represents the
line of intersections of all logarithms with their corre-
nAnlegeitins
ey
swine Figure 30
sponding antilogarithms.
As the logarithm increases at a steady rate (0.000,
0.301, 0,602, 0.903, 1.204), while the antilogarithm in-
creases at a steadily increasing rate (I, 2, 4, 8, 16), the
curve bends upward, becoming ever more nearly vertical.
If the horizontal lines, representing successive antiloga-
rithms from 1 to 10, are reflected, so to speak, upward
from the curve, we get a sequence of numbers from 1 to
10 at the top of the graph that are squeezed to the right
exactly as they are on the D-scale. The antilogarithms,
as positioned at the top of the graph, are exactly over
their corresponding logarithms at the bottom of the
graph. The top of the graph represents the D-scale,
therefore, and the bottom of the graph the L-scale.
Since the compression at the right in the D-scale is
brought about by the adjustment of its numbers to a
scale of logarithms increasing at a fixed rate, the D-scale
is an example of a logarithmic scale.
Subdividing the D-scale
The spaces between the primasies on the D-seale can
be divided into finer subdivisions. Consider the space
between D-1 and D-2, for instance. D-1 is under L-000
and D-2 is under L-30) because log 1 = 0 and log 2=
0.301, Using the same system we can place the values
11, 1.2, 18, and so on, onto the D-scale directly over
appropriate values on the L-scale.
Thus, log 1.1 = 0.041; log 1.2 = 0.079; log 1.3 = 0.114,
ete, Therefore, D-1.1 can be placed over L-041; D-1.2
over L-079; D-1.3 over L-114, and so on (Figure 31).
‘This gives us our secondary markings between D-1 and
D-2, Similarly, tertiaries can be placed between D-1.1
and D-1,2, marking off-1.11, 1.12, 1.13, and so on. In
this way, too, secondaries and tertiaries can be set up
all along the length of the D-scale as finely and as ac-
curately as the length of the slide rule and the delicacy ofthe instruments used will allow.*
The usual secondaries and tertiaries on the D-scale off
an ordinary slide rule are shown in Figure 32. You will
notice two things. First, the secondaries and tertiaries
get more crowded as one moves to the right. This is to be
expected in a logarithmic scale. The second is the con-
sequence of this, for the tertiaries represent different,
values in different portions of the scale.
‘This should not be surprising. In an ordinary seal
such as the L-scale, the spacing between the pr e
is equal throughout and secondaries and tertiaries can
be distributed evenly all along the line. In a logarithmic
* The difficulty of making an accurate logarithmic scale com
tributes to the expense of a good slide rule.
Figure 32
naynonmepg
Figure 31
scale, however, the space between the primaries steadily
decreases as we move to the right so that room can be
found for fewer and fewer subdivisions.
‘As it happens, secondaries representing 0.1 each are
distributed all along the D-scale but get progressively
closer, and the space between them available for ter-
tiaries gets progressively skimpier. Consequently, there
are tertiaries of three different kinds present on the
D-scale,
Consider first, the space between D-1 and D-2, which
takes up nearly a third of the total length of the D-scale.
It can be divided up finely. The secondaries, represent-
ing tenths, are so relatively far apart that it is hard for
the eye to take them in at a glance. For that reason (and
ong nomeFigure 33
because there is enough room for it), the secondaries
between D-1 and D-2 are marked with the numbers 1
through 9, numbers of smaller size than those marking
the primaries (see Figure 82). The small number 1
marks the 1.1 value, the 2 marks the 1.2 value, and so on
up to the 9, which marks the 1.9 value.
The spaces between these secondaries is divided into
ten smaller divisions. Thus, there are tertiaries, each
representing an increase in value of 0.1, running from
1.01 for the first tertiary after D-1 to 1.99 for the last
tertiary just before D-2. The middle tertiary between
neighboring secondaries, marking numbers such as 1.05,
1.15, 1.25, and so on, is longer than the others.
In addition, one can place the hairline midveay be-
tween two tertiaries so that one can indicate 1.005 or
1.235 without trouble, or make rough estimates for 1.007
or 1.288.
Beyond D-2, the secondaries do not possess numbers
actually marked on the scale, If numbers were present
they would be crowded too closely for clarity. On the
other hand, the middle secondary in each case, repre-
senting 2.5, 3.5, 4.5, and so on, is made longer than the
rest, and using these as guides, the value of any other
secondary is quickly seen.
16
erasing 116
genre
$$
erin = 008
Since the stretch from D-2 to D-4 is just equal to that
from D-1 to D-2, there are twenty secondaries crowded
into the former stretch, where only ten exist in the latter.
For that reason, there is no room in the D-2 to D-4
stretch for the secondaries to be conveniently divided into
ten smaller divisions, Five smaller divisions are set up
instead and each tertiary in this stretch therefore repre-
sents a value of 0.02. Thus, between D-2.1 and D-2.2
are tertiaries representing 2.12, 2.14, 2.16, and 2.18, By
centering the hairline between such tertiaries, one can
get 2.18, 2.15, and so on.*
The stretch from D-4 to D-10 is divided into sixty
secondaries that crowd closer and closer together. So
close are they that each space between secondaries in
this portion of the D-Scale is divided merely into two
smaller divisions, with a single tertiary representing a
value of 0.05, Thus, between D-4.3 and D-4.4 is a tertiary
indicating 4.35, and between 9.6 and 9.7 is a tertiary in-
dicating 9.65. By centering the indicator between a
tertiary and a secondary, one can obtain numbers such
as-4.825 or 9.675.To summarize, we can divide the D-scale into three
segments in which the value of the tertiaries differs (see
Figure 33). From D-1 to D-2, each tertiary represents a
value of 0.01; from D-2 to D-4, a value of 0.02; and
from D-4 to D-10, a value of 0.05.
Naturally, when one is not accustomed to a logarithmic
scale, this change in the nature of the tertiaries can
be confusing. There will be some stumbling at first,
inevitably, but continued use of the slide rule will
eventually make it all seem natural and there will be no
trouble in reading the D-scale, no more than in reading
the markings of an ordinary ruler.
It might seem to you from all this that the left end of
the D-scale is more delicate than the right end, since
the left end has tertiaries that represent finer divisions.
Let’s look into that situation.
Suppose you are trying to read the hairline position on
the left end of the scale, and find it just about halfway
between the tertiaries 134 and 1.35. You conclude that:
the reading, therefore, is 1.345. However, it may seem to:
you that the position is just a shade to the right of the
midpoint, so that the reading might be 1.346, You can't
really tell exactly which of these it is so you have an un-
certainty of 0.001.
‘On the right-hand side of the scale you might make a
reading of 9.475, because the hairline is just halfway
between the tertiary marking 9.45 and the secondary
marking 9,50. However, again you cannot be sure that
it is cxactly in the center. The space between 9.45 and
9.50 (a difference of 0.05) is only about two-thirds as
wide as that between 1.34 and 1.35 (a difference of only
0.01) and its exact midpoint is harder to judge. If the
18
hairline seems a bit to the right of the midpoint, the
reading can easily be 9.485 rather than 9.475, an un-
certainty of 0.01.
Since the uncertainty at the left is 0.001 and the uncer-
tainty at the right is 0.01, you might think that the
accuracy at the left end is ten times as great as that at
the right end. That, however, is not so.
‘What counts is not the size of the uncertainty in itself,
but its size compared to the size of the number being
measured. What we ought to consider is the uncertainty
percentage. Thus, if you consider an uncertainty of 0.001
in a number such as 1.845, the uncertainty is about 0.075
percent of the number. An uncertainty of 0.01 in a
number like 9.47 represents an uncertainty that is about
0.105 percent of the number.
These percentages are not very different. We can
say, with reasonable confidence, that the accuracy of
a slide rule is the same all along its length and that it is
everywhere easily reliable to about a tenth of one per-
cent, or one part in a thousand. (A very careful user
could squeeze out an even better accuracy.)
Using the Wooden Log Table
Let's consider the L-scale and the D-scale taken
together. What can be done with them? Obviously, the
two scales have been constructed in such a way as to
give us a three-place logarithm table, with the logarithms
in the L-seale and the antilogarithms in the D-scale,
Suppose you want the logarithm of 3.5. You place
the hairline over D-3.5 and find that it also marks L-544
(Figure $4). You can conclude then that log 3.5 = 0.544.
ryFigure 34
If you were to look up the logarithm of 8.5 in a five-place;
logarithm table, you would find that log 3.5 = 0.54407 so;
that your slide rule answer is certainly satisfactorily close,
particularly since all you need to do to get it on the slide
tule is to shift the hairline. It is a much simpler task
to do that than to leaf through an ordinary logarithm.
table and run your finger down columns of numbers.
Of course, the L-scale gives you only the mantissa, |
but that is all you need, If you wanted the logarithm ofj
35, or 350, or of 0.35, you would still place the hairline|
on D-8.5 to get L-544. You then merely adjust the char-
acteristic according to the principles described on page
54. We then see that log 85 = 1.544, log 950 = 2.544,
log 0.35 = 0.544 — 1, and so on.
It works the other way round, too. Suppose you are
given the logarithm 8.728 and want the antilogarithm,
To use the L-scale, which gives only the mantissa, you!
drop the characteristic temporarily and search for 728}
only. You place the hairline at L-728 and find that the!
D-reading is just a shade to the left of 5.35 (see Figure
84). You estimate the reading to be 5.345. The charac~
teristic tells you where the pe ought to go. Since 4
cprmceror nnn,
Aatiog 276 = S96
the characteristic is 3, the number of digits to the left of
the decimal point must be 3 + 1, or 4. Consequently,
antilog 3.728 = 5845. A five-place logarithm table will
tell you that antilog 8.728 = 5345.4.
As you see, then, the D-reading is not fixed. What
we read as 3.5 can be 350; and what we read as 5.345 can
be 5345 or 0.0005345, for that matter. A reading on
the Descale gives us a fixed set of digits, but it says
nothing about the position of the decimal point. That
belongs to us.
For this reason, I will no longer try to place a decimal
point in the D-reading. At first, it may seem very con-
venient to consider the secondary marks as representing
“tenths” and to read them as 1.1, or 2.4 or 9.6, but 1
will consider such readings as D-11, D-24, or D-86 re-
spectively, In the same way tertiary marks will not
represent 1.46 or 2.46 or 9.55. They will be D-146,
D-246, and D-955 respectively.
You may feel lost without the decimal point just at
first, hut you will get used to that too. And the loss is
only temporary. As you will see, once a slide rule compu-
tation is completed, in will go that decimal point.5
Multiplication
The C-scale
WE HAVE now converted the slide rule into a logarithm
table and, having done so, we find that we don’t really
need to labor in order to use it. The slide rule makes its
use automatic. Whenever the hairline is placed on any
D-reading, it simultaneously marks off the logarithm of
that number as the L-reading. We don’t have to leok
for the logarithm; it is there.
This helps us understand a crucial difference between.
the evenly spaced scale on an ordinary ruler (such as
the scales we used to work our addition rule in Chapter
2) and the logarithmic D-scale.
In an ordinary scale, the numbers mark off lengths
that correspond in value'to the numbers themselves. The
number 2 is two inches from the left end; the number 5
is five inches from the left end and so on. By adding and
subtracting these lengths, therefore, we add and subtract
numbers (see pagell).
The numbers on the J)-scale, however, mark off lengths
that correspond not to the numbers themselves, but to
the logarithms of the numbers. On the L-scale, the
primaries are usually placed about an inch apart (see
page 62). That means that since D-1 is placed over L-0
82
and D-2 is placed over L-801, D-2 is about 3 inches from
the left end of the scale. Similarly, since D-3 is under
L-477, D-3 is about 4% inches from the left end of the
scale. These distances, I repeat, are not equivalent to
the numbers themselves but to the logarithms of the
numbers.
Consider, then! When we used an addition rule
with its scales representing lengths equivalent to the
numbers upon it, we could manipulate those lengths
so as to add and subtract numbers. But if we use some-
thing like the D-scale in which the numbers represent
lengths equivalent to the logarithms of those numbers,
we can manipulate those lengths so as to add and sub-
tract logarithms. In adding and subtracting logarithms,
we are, of course, multiplying and dividing the anti-
logarithms — that is, the numbers on the D-scale.
To perform such manipulations on the addition rule,
we needed two scales of identical construction that could
he moved against each other. We need the same now
— two scales of identical construction that can be moved
against each other.
Therefore, on the bottom of the slide, in the C-position
(see page 66 ), another scale, exactly like the D-scale, is
constructed. This new scale, the C-scale, can be placed
exactly over the D-scale, with every marking on the
former over the corresponding marking on the latter and
we can then say the slide rule is in a neutral position
(Figure 35). The C-scale can, however, be shifted out
of neutral and moved along the D-scale in such a way
as to add and subtract logarithms,
‘We can begin with a very simple: 2 X 8.
To do this, we follow a similar procedure to that usedJnana
ee
pte ft fanaa
Figure 35
in adding numbers on the addition rule (see page 19 ).
We begin by moving C-1 over D-2.
And this at once brings up an important point. There
are two C-I's, one at the left end of the scale and one at
the right end. These are used so often in slide rule
manipulations that they are often referred to as the left
index and the right index, respectively. In this book,
however, in order to avoid confusion, I will give no
readings any special name. I will refer to the 1 at the
Jeft end of the C-seale (the more frequently used) as C-1.
The 1 at the right end of the C-scale, I will refer to as C-1-
(right). In the same way there is both a D-1 and a D-1-
(right).
In placing C-1 over any D-reading, it is not necessary
to use the hairline as a guide. Standing as it does at the
very beginning of the scale, C-1 is clearly marked out
and the eye can follow it without trouble. Markings
along the interior of the scale are easy to confuse with
their neighbors, however. Readings between markings
are even easier to confuse. It is routine, then, to place
the hairline.over any reading one wishes to make that
does not involve the end-marking of a scale.
With all this in mind, we can return to the multiplica~
tion 2X 3. Carrying it thrayeh by the system we used
in the addition rule, we place C-1 over D-2, then shift the
hairline over to C-3 (Figure 36).
Since D-2 marks off a distance of log 2 from the end
of the D-scale, and C-3 marks off a distance of log 8 from
the left end of the C-scale, we are adding log 2 and
log 3 in this manipulation. Since log 2+ log 8 = log 6
(and you can check this in a log table), we would expect
to find D-6 under C-3, for D-6 would be a distance of
log 6 from the end of the D-scale.
We do not, however, bother reading the slide rule in
the position shown in Figure 86 as an addition of loga-
rithms. We do not say: log 2 + log 3=log 6. The D-
scale gives the antilogarithms and we read those anti-
logarithms directly, converting the addition of loga-
rithms into a multiplication of antilogarithms. We say,
2x3=6.
‘The same thing would happen if we wanted to mul-
tiply 3 X 2. We would place C-1 over D-3, then move
the hairline to C-2, and find D-6 immediately under C-2.
This is not surprising for log 2 + log 3 = log 3 + log 2,
and 2X8=3X2
This is simple so far. We scarcely need a slide rule to
tell us that 2X 8 = 6. However, what if it is 2.54 x 3.76
that we are interested aol ad multiplication is nowmuch more difficult on paper, but not a bit more difficult
on the slide rule.
Move C-1 over D-254 (taking a little care to make
sure you have the right tertiary), and then move the hair-
line to C-376, You find the hairline to be over D-955 (see
Figure 37) and conclude that 2.54 x 8,76 = 9.55. Work
it out in full and you will find that the correct product is
9.5504, but surely 9.55 is close enough — and think of
the saving in time.
But what if it were 2.54 X 4.76 you wanted to solve?
You begin, again, by moving C-1 over D-254, but this
time you find you are stuck, for the C-476 which you
must reach with your hairline is off the D-scale and can
give you no D-reading.
Ah, but we went through that in the addition rule
(see page 21) and the same device will serve us here.
We make use of the C-1(right) in place of the C-1 and
put that over D-254. Now we can find C-476 over the
D-scale. The nearest tertiary is C-475 so we place the
hairline just-a bit to the right of that, about one fifth of
the way over to C-480. (Naturally, this means a kind of
“judging by eye” but this can be done pretty well,
especially with practice.) The hairline, on C-476, also
falls just to the left of D-121 (which can also be read
as D-1210). We might judge the mark to be D-1209
(Figure 88), and conclude that 2.54 X 4.76 = 12.09. Ifase na
we work the multiplication in full, the product turns
out to be 12.0904.
The Decimal Point
This brings up the question of the decimal point. In
the example just completed (2.54 X 4.76) the slide rule
gave us the digit-combination 1209 as the answer. We
might have decided that this meant 2.54 X 4.76 = 1.209,
or that 2.54 X 4.76 = 120.9, but we didn’t. We said
that 2.54 x 4.76 = 12.09, Let's see why.
Consider the following multiplications:
15 X 823 = 4845
15 x $23 484.5
0.015 x 0.323, 0.004845,
. 150,000 0.00828 = 4845
} 150 XX 323 48,450
‘This shows us once again that the decimal point is a
side issue. If we go by the slide rule we deal only with
digit-combinations and all the examples above boil down
Figure 38
to 15 X 823 = 4845. The decimal point — that side issue
—is for us to handle.
There is, however, no need to feel aggrieved, for,
though placing the decimal point may be a tedious ne-
cessity, it is not difficult. Occasionally, a bit of careless-
ness will result in a misplaced decimal point, but this can
happen even in pencil-and-paper calculations, The cor-
rect response to carelessness is a sober determination to
be careful, that's all!
‘The best way to place the decimal point is to consider
the problem and substitute, for the numbers involved,
similar numbers that are particularly easy to handle.
Such similar numbers will give you an answer that is
wrong, of course, but one that is close enough to the
right answer to have the decimal point in the same place.
You will have an answer that is of the same order of
magnitude.
Each shift of a decimal point by one place produces a
change of one order of magnitude. For instance, 2.54 is
one order of magnitude smaller than 25.4 and two orders
of magnitude smaller than 254, Again, 2.54 is one order
89of magnitude larger than 0.254 and two orders of magni-
tude larger than 0.0254,
If two numbers differ by a factor of less than 5, they
can be Jumped together as of the same order of magni-
tude. Thus, 17 and 68 are of the same order of magnitude
since 68 is only four times as great as 17. On the other
hand 17 is not of the same order of magnitude as 2.14
since 17 is eight times as great as 2.14. However, 17 is
of the same order of magnitude as 21.4 which is, in turn,
‘one order of magnitude greater than 2.14. Therefore,
we can say that 17 is one order of magnitude greater
than 2.14,
Now let's look at one of the problems in the list given
‘on page 88, say, 15 X 82.8. We can convert 15 to 20
and 32.3 to 30. The numbers are changed but not the
order of magnitude. It is easy to multiply 20 by 30 in
our head. The answer is 600. That is not the answer we
are looking for, but it is the same order of magnitude as
the answer. If the slide rule tells us that the digit-combi-
nation of the product is 4845, then to make that the same
order of magnitude as 600, we must write it 484.5. We
know then that 15 X 82.3 = 4845.
A similar treatment will give us the correct answer for
any other example in the list.
‘And it will also give us the correct answer to the
problem with which I started this section, 2.54 x 4.76,
If we make the first number a little higher, changing it to
8, and the second a little lower, changing it to 4, we
retain the order of magnitude. Since we know that
3X 4 = 12, we know that the slide rule answer of 1200
must be written 12.09 to keep the product in the correct
order of magnitude. Therefore, 2.54 X 4.76 = 12.09 and
nothing else.
Practice is all you need to learn to handle orders of
magnitude almost automatically, and once you've worked
that out in advance, you can manipulate the slide rule, get
your digit-combination, and place your decimat point
without any hesitation.
Let's take another example, and a slightly more difficult
one.
Suppose you wanted to carry through the follow:
2.72 X 7.23 X 1.15 X 0.86. This represents three nulti-
plications to be carried out one after the other.
You can begin by placing C-1 over D-272 hut then you
‘will find that C-723 is off the D-scale. You therefore
“switch indices,” bringing C-1(right) over D-272, and
carrying the hairline to C-723. (To find C-723, you must
place the hairline between the secondary representing
720 and the tertiary representing 725 — and place it a
little closer to the 725 than to the 720.)
Having done this, you will find that the hairline is also
marking out a reading just under D-1970. You don't,
however, have to try to estimate what the answer is, or
even look at it. The hairline marks the product of
2.72 X 7.23 on the D-scale and keeps marking it as long
you leave the hairline in place, This product must next
he multiplied by 1.15.
You therefore bring the C-1 to the hairline (making
sure you don’t move the hairline position in the process )
and this places C-1 over the D-reading that is the product
of 2.72 x 7.23. To multiply that product by 1.15, you
move the hairline (which you no longer need in its first
position now that the C-1 marks it) to C-115. That marks
a new D-reading which represents the product of 2.72
a7.23 X LIS.
This new product must be multiplied by 0.86. You
therefore bring C-1 to the hairline marking that product
and find that C-86 is off the D-scale. You bring C-1-
(right) to the hairline instead and then move the hairline
to C-86.
Now the hairline marks out a D-reading just about mid-
way between D-194 and D-195. Call it D-1945, That is
the digit-combination that represents the product of
2.72 X 7.23 X 1.15 X 0.86 (Figure 39)
But where is the decimal point? If we look at the
multiplication problem again, we can change 2.72 to 8,
7.23 to7, 1.15 to 1, and 0.86 to 1. The problem becomes
3X 7X 1% 1 which you can see at a glance equals 21.
To give 1945 the same order of magnitude as 21, we
must write 1945 as 19.45. Therefore 2.72 X 7.28 X 1.15
X 0.86 = 19.45,
(Ihave, by the way, just worked out the triple multipli-
cation by pencil and paper as quickly as I could and it
took me two full minutes to get the answer 19.4492784. I
worked it out slowly and carefully by slide rule and it
took me just 15 seconds — one-eighth the time — to get
19.45, which is almost exactly the correct answer.)
Folded Scales
Sometimes, as you saw in the case discussed just above,
you must use C-1 and sometimes C-1(right). With
practice, you get the “feel” for which one to use.
Suppose, for instance, you consider 2.14 X 87.6. You
can adjust these figures to their easy approximations of 2
and 40 and see, at a glance, that 2X 40 = 80. In that
92
case, you will recognize that it is safe to bring C-1 to
D-214 for you expect C-376 to be somewhere near D-8,
which would keep it safely on the D-seale. Sure enough,
if you perform the manipulation, you will find C-876 over
‘1D-805 so that 2.14 X 37.6 = 80.5.
If, on the other hand, you were dealing with 4.14 x
3.76, which is close to 4 X 4, which in turn is equal to 16,
you would expect the answer to be beyond D-1 (right),
which would represent a digit-combination of only 10.
You can consider D-1 itself to be 10 also, of course, but
in that case 16 would be near the left end of the D-scale
and must be sought there. Therefore you bring C-1-
(right) to D-414 and then find C-876 over D-1556 so that
414 X 8.76 = 15.56.
However, it is always possible to be fooled. Suppose
you try 13X79 and decide to approximate it at 10 X
80 = 800, Therefore you move C-1 over to D-13 expect-
ing to find C-79 somewhere near D-8 and safely on scale.
You find instead, to your horror, that C-79 is just beyond
the end of the D-scale. You must therefore switch indices
and bring C-1(right) all the way over to D-18 and then
find C-79 over D-1027 (Figure 40) and decide that
13 X 79 = 1027.
‘Whenever this happens, there is bound to be a certain
amount of muttering under the breath.
It would be nice if this could be avoided. Suppose, for
instance, that the D-scale were continued past D-1-
(right) so that there were a D-2(right), a D-3(right)
and so on, Then, if a C-rcading moved off the ordinary
D-scale, it could be picked up on the D-seale extension.
The disadvantage of this is that the slide rule must be
lengthened, making it both more expensive and morepanxm) oun cm «us
syslog
unwieldy. (In some stide rule designs, however, par-
ticular scales are extended about half an inch or so past
either end to take care of Toggertins cases.)
Figure 39
Another way out is to move the D-scale bodily to the
left so that D-1(right) is moved leftward, leaving room
beyond it for an extension i making the slide ruleunr on
any longer in that direction.
If you keep the rest of the D-scale intact, however, it
then begins to stick out at the left end like a tusk. In
order not to extend the slide rule in either direction, we
must arrange to have the left end of the D-scale disappear
as it moves leftward. The additional room at the right for
an extension must then be made up for by a disap-
pearance of the scale at the left (Figure 41).
As you move D-I(right) leftward — provided you |
keep the overall length of the scale fixed — the extension
you gain on the right is exactly balanced by the loss on
the left. It follows then, as you can see in Figure 41, that
the reading on the left end of the scale is always the same
as on the right end of the scale. If the scale begins with
D-2, it ends with D-2(right); if it begins with D-4 or D-8,
it ends with D-4(right) or D-8(right).
‘As you move D-1(right) leftward, you experience a
gain because of the extension on the right, and you also
experience a loss because of the disappearance on the
left. To increase the gain as much as possible you want
to push D-1(right) continually leftward. To minimize
96
Figure 40
the loss you want to push D-1(right) continually right-
ward, The two impulses balance, as you might expect,
precisely midway and the ideal situation is to have D-1
exactly in the middle of the scale.
If D-1 were exactly in the middle of the scale what
would be the numbers on the left end and right end?
(Whatever number would be at the left end would, ofEom pcerpte anf
course, also be at the right end.)
To answer this question, let us remember that the
entire stretch of the D-scale represents the stretch of the
L-seale from 0.000 to 1.000 (see page 67). If D-1 is in
the iniddle of the scale, the logarithmic stretch from that
middle to the right end is, obviously, 0.500 or, in fractions,
4, Earlier in the book, I pointed out that 10% = V/T0, or
3.16. Therefore, antilog 0.500 = 3.16.
‘The numbers on the D-scale are placed at the distance
from D-1 that corresponds to the length of the logarithm
of those numbers. Therefore the right end of a D-scale
that has D-1 in the middle must be 3.16, since the
distance of 3.16 from D-1 is then just half the length of
the scale or 0.500 (in logarithms ) and that is, indeed, the
logarithm of 3.16. Consequently the left end of the D-
scale with D-1 in the middle must be 8.16 also. Omitting
decimal points, a D-scale D-1 in the middle should
run from D-316 to D-816( right).
Such a scale is called a folded scale, the notion being
that an ordinary scale, folded in half, has 316 appearing
at the fold, and the new seale starts and ends at that fold,
In my slide rule, the folded scales are present in the A-
and B-positions (see page66), but they are not referred
to as A- and B-scales. Those names are reserved for
other scales which I will aiscuss later in the book.
inva} fron of sss
Tovntunduunefiant fend oH ome on avant woo
Figure 42
Instead, the scale in the A-position is the DF-scale,
while the scale in the B-position is the CF-scale (Figure
42). The DF-scale is so named because, like the D-scale,
it is on the body, while the CF-scale, like the C-scale, is
on the slide. The F, of course, stands for “folded.”
Pi
If you look at Figure 42 carefully (or, better still, look
at a slide rule) you will see that the DF-scale and the
CF-scale do not actually start and end with 3.16 but
with 9.14 and that the two ends are marked thus: 7.
Behind that lies a story. In the case of every circle,
however large or small, the length of the circumference is
about imes that of the diameter. The actual ratio is
a never-ending decimal, but, to five decimal points, itis
3.14159 which, as you see, is not far removed from 3.14.
‘This ratio is customarily symbolized by the Greek letter
= (“pi”).
Tt-so happens that x is not confined merely to questions
involving the length of the circumference of a circle in
terms of its diameter. It crops up in numerous mathe-
matical equations and comes into play in almost every
facet of science.
Consequently, scientists and engineers frequently findoccasion to multiply « by some number. That is the
reason for the special x marking on the C- and D-scales.
It is also the reason why the DF- and CF-scales are de-
signed to start with x. Although those scales should,
ideally, start with 3.16, it is not much of a departure from
the ideal to have them start just to the right of 3.14
instead. It is an alteration of not quite a single tertiary
and the insignificant bit of imperfection introduces such
a great deal of convenience that no one would object.
To see what this means, let's begin by noting that C-1 is
directly under CF-x. This is a fixed position, for both the
C-scale and the CF-scale are on the slide and neither can
be moved relative to the other. This is also true for the
D-scale and DF-scale which are both on the body so that
‘one can't be moved relative to the other. Therefore, as-
suming both parts of the body to be in perfect adjustment,
D-1 is permanently under DF-
Next, let’s consider the hairline at some point on the
C-scale — any point. Let us say the reading is C-a. At
the same time, the hairline is indicating a reading on the
CF-scale; say, CF-b.
We already know that the C-a reading is at a distance
from the left end of the C-scale equivalent to the log-
arithm of a. Since the left end of the C-scale is C-1, which
represents a logarithin of 0.000, the logarithm of the C-a
reading is 0,000 + log a, or simply log a.
‘The CF-b which is simultaneously marked off by the
hairline also represents a logarithm, which we can call log
b, But CF-b (marked out by the hairline simultaneously
with C-a) must be as far from the left end of the CF-scale
as C-a is from the C-seale. Hence CF-b is a distance from
CE-« that is equivalent to log a. But CF-r is equivalent
to log . Therefore, log b=log x + log a. If we convert
those logarithms into antilogarithms (and remember to
change the addition into a multiplication in so doing), we
find that b = x X a ( Figure 43).
If this is not instantly clear to you, it will become clear
as soon as you put the notion into actual practice.
Suppose you want the value of 2n. You have only to
move the hairline to C-2, and you will find the hairline
will simultaneously give a reading of CF-6.28, which is
equal to 2r. .
In the same way you can find with a single setting
of the hairline that 4.3r = 13.51, since, when the hair-line indicates C-43, it also indicates CF-1951. As
other examples you can find that 75.2 7 = 236, that
0.388r = 1.22, and so on. The placing of the decimal
point is no problem, for you have but to remember that
multiplying by x is not much different from multiplying,
by 3.
Avoiding the Shift
But multiplication by = is a side issue. When I started
discussing the folded scales, it was not with such a multi-
plication in mind. The folded seales were introduced in
order to make it unnecessary to shift from C-1 to C-1-
(right) in ordinary multiplications. (Or at least to reduce
the number of times the shift must be carried through,
if the necessity isn’t eliminated altogether. )
In order to go back to this earlier problem, take another
Figure 46,
ily
ut
Figure 44
look at the CF/DF scales as compared with the C/D
scales. When the slide rule is in its neutral position, C-1
is over D-1, and CF-1 is under DF-1, as shown in Figure
42.
If the slide now moves in such a way as to displace the
C-scale aganst the D-scale, it also displaces the CF-scale
against the DF-scale. And since the C-scale and the CF-
scale are on the same piece of wood, the amount of dis-
placement must be precisely the same for both.
Suppose, for instance, that you move C-1 over D-226.
You will find that CF-1 will move directly under DF-226
(Figure 44). Furthermore, if you look at other portions
of the scales, you will see that there are similar situations
all along the line. Thus, C-13 is directly over D-204 and
CF-13 is directly under DF-294, and so on for any other
readings you care to make,
Since the readings on the C- and D-scales are exactly
oregano
nada torres orgaduplicated by the readings on the CF- and DF-scales, it
doesn’t matter which pair of scales you use for multiply-
ing. You can use the folded scales exclusively if you wish;
‘or you can move freely from the ordinary scales to the
folded scales and back again.
‘There is an advantage to moving back and forth be-
tween the two sets of scales. Suppose C-1 is placed over
D-314. In that case, half the C-scale moves beyond the
right end of the slide rule and the markings from C-314
into the higher numbers can’t be used. Half the CF-scale
also moves beyond the right end of the slide rule, but here
it is the numbers smaller than CF-314 that can’t be used
(Figure 45).
The situation is, then, that numbers up to 314 are
usable on the C-scale and numbers over 314 are usable on
Figure 47
the CF-scale. On one scale or the other, all the numbers
are usable even though half the slide is beyond the right
end of the slide rule. If, then, in adjusting either C-1 or
‘C-1(right) over a D-reading, you are careful to move the
slide less than half the length of the slide rule you can be
assured that all possible multiplications can be made
without shifting indices,
Suppose, for instance, you wanted to multiply 2.3 by
4.8. Move C-1 over D-28 and you would find that C-48
was beyond the right-hand end of the D-scale and there-
fore unusable. One way out would be to heave a sigh and
shift indices so that C-1(right) is over D-23. This, how-
ever, is not necessary. Leave C-1 over D-23 and simply
look over to the CF-scale. You can find CF-48 with no
trouble at all. It is directly under a point between DF-110 and DF-111, somewhat closer to DF-110 (Figure
46). Call it DF-1104, therefore, and you conclude that
2.3 X 4.8 = 11.04
In the same way, if you wish to multiply 5.4 by 1.45
and move C-1(right) over D-54, you will find that C-145,
is beyond the left end of the slide rule. Calmly, you
switch to CF-145 and find it just to the left of DF-785
(Figure 47). You judge the reading to be DF-783 and
decide that 5.4 X 1.45 = 7,83.
‘The usefulness of the folded scales is not confined only
to those times when you have misjudged and used the
wrong end of the C-scale. Suppose (as frequently hap-
pens) that you must multiply a whole series of numbers
by a particular factor. For instance, there are just about
2.2 pounds in a kilogram and you may be given a number
of weights, in kilograms, which you wish to change into
pounds, Each weight in kilograms must be multiplied
by 2.2 if that purpose is to be accomplished. Suppose that
the kilogram weights given you are 12.3, 32.1, 46.7, and
743.
You move C-1 over D-22 and leave it there. That
represents a general multiplication of any number by
2,2. Now you move the hairline to C-123 and find it
over D-271; you move the hairline to C-321 and find it
over D-706. So far, so good; 12.8 kilograms = 27.1
pounds and 82.1 kilograms = 70.6 pounds.
But C-467 and C-473 are beyond the right end of the
D-scale. No bother! Don't shift indices! Just move over
to the CF-scale where you will find CF-467 under
DF-1027 and CF-743 under DF-1635. Therefore, 46.7
kilograms = 102.7 pounds and 74.8 kilograms = 163.5)
pounds,
106
Division
Reversing Multiplication
Division is the reverse of multiplication, as subtraction
is the reverse of addition. When I described the addition
tule, I pointed out that we could perform subtraction by
running the technique for addition in reverse (see page
81). On the slide rule, we can carry out division by
running the technique for multiplication in reverse.
Tn multiplying 2 by 3, we place C-1 over D-2 and look
under C-3 to find D-6. Therefore 2 x 8 = 6.
In dividing 6 by 3, we place C-8 over D-6* and look
under C-1 to find D-2 (see Figure 48). Therefore
6+8=2. The reason for this is not hard to see. The
position of D-6 represents a distance of log 6 from the left
end of the D-scale. The position of C-3 represents a dis-
tance of log 3 from the left end of the C-scale. If we place
C-3 over D-6, and follow the C-scale back to C-1, you see
from Figure 48 that we are subtracting log 3 from the
‘In setting « C-reading over a D-teading, where neither reading
ide.
‘then bring C-3 to the haidine. The hairline (s particularly useful
‘where the readings are not on actual markings. To place C-319
over D-413, for instance, without using the hairline, is extremely
dificult, 23 you can assure yourself if you have a slide rule
hundy. With a hairline, it requires 2 little elose attention, but
isnot particularly hard.te
distance of log 6 on the D-scale. What is left of the D-
scale beyond C-I is log 6— log 8. Since the reading
under C-1 is D-2 (which is a distance of log 2 from the
left end of the D-scale), we can say that log 6 — log 3=
log 2. But subtracting logarithms is equivalent to divid-
ing antilogarithms; therefore 6 + 3 = 2.
‘Any division can be carried through in the same
fashion. If you want to divide 48.5 by 7.4, you place
‘C-74 over D-485 and you will find that C-1 is off the scale.
That needn't bother you, for you need only look under
C-1 (right) instead and find D-655 (see Figure 49). You
decide therefore that 48.5 + 7.4 = 6.55.
How did you determine the position of the decimal
point? In the usual way. Instead of 48.5 + 7.4, you make
sna omer niga otb os
Figure 48
use of the near-equivalent 49 = 7, to which the answer,
as you see at once, is 7. This is not the right answer, but
it is the right order of magnitude,
In dividing you place a C-reading over a D-reading
and then look for either C-1 or C-1(right). No matter
how you adjust the slide, either C-1 or C-1(right) will be
on the D-scale so that you can always find a quotient
without reference to the folded scales.
Nevertheless the folded scales do come in handy for the
special case of division by x. For that purpose we can
make use of the CF-scale, reversing the procedure for
multiplication of x (see page 118). Since the C-reading
multiplied by equals the-CF-reading, the CF-reading
divided by = equals the C-reading. If you want theanswer to 6 + n, you have only to set the hairline on CF-6 In the same way, you can determine that 25.4 + «=
and it will simultaneously read C-192 (see Figure 50). 8.09, and so on.
Since x is approximately 3 and % = 2, you have to set There is one possibility of confusion in division that
your order of magnitude and can say that 6 + = = 1.92. does not arise in multiplication. The order in which you
edna igang sent fant,multiply does not matter; so that 7.62 X 1.44 will give the
same product as 1.44 X 7.62. Therefore it doesn't matter
whether you put C-1(right) over D-762 and move the
hairline to C-144, or put C-1 (right) over D-144 and move
the hairline to C-762. In either case, you will end with a
reading of D-110 and you will decide that 7.62 X 1.44 =
110.
The order in which you divide, however, does matter,
as it does when you subtract (see page 84). You can see
that the quotient of 75+ 48 is not the same as that of
48 + 75. In the former case, you put C-48 over D-75 and
find D-156 under C-1. In the latter case, you put C-75
over D-48 and find D-64 under C-1(right) (Figure 51).
Since 75 + 48 can be replaced by the near-equivalent
75 + 50, which equals 1.5, and 48 + 75 can similarly be
replaced by 50 ~ 75 = 0.67, we see what our order of
magnitude is and where to place the decimal point. We
decide, then, that 75 = 48 = 1.56 and 48 + 75 = 0.64.
But how do you remember which number goes on the
C-scale and which on the D-scale? When do you put
C-48 over D-75 and when C-75 over D-48? Actually, the
decision is a simple one.
In any division, as, for example, @ + b, ais the number
being divided into and is the dividend, while b is the
number doing the dividing and is the divisor. The rule,
then, is that the divisor goes on the C-scale and the
dividend on the D-scale. (This is analogous to the man-
ner in which the subtrahend — analogous to the divisor
—is placed on the slide in the addition rule. See
page 38.)
In the case of 75 + 48, 48 is the divisor and so C-48
goes over D-75, In the case of 48 ~ 75, it is 75 that is the
2
divisor, and it is C-75 that goes over D-48. When this
rule is followed, the quotient is always found on the
D-seale undemeath whichever C-1 happens to be on
scale.
But what if, by mistake or even by intention, the divisor
is placed on the D-scale. It is not good to place matters
in reverse thus, for if you follow one set rule, it will he-
come automatic make the use of the slide rule that
much less of a conscious effort. Still, if you do reverse
the rule, all is not lost.
Suppose we take the simple example 42 + 7 and try
it first in the ordinary fashion. We place C-7 (the
divisor) over D-42 (the dividend) and directly under
C-1(right), we find D-6, Therefore, 42 +7 = 6.
Now we'll try it “upside down.” We will place C-42
over D-7 and we must now tum the procedure for finding
the quotient upside down as well. Instead of finding
the quotient on the D-scale under whichever C-1 is on
scale, we find it on the C-scale over whichever D-1 is on
scale. In this case, D-1(right) is under C-6 and again
42+7=6 (Figure 52).
Whichever way you carry through your division —
with the divisor on the C-scale and the dividend on the
D-seale, or the divisor on the D-scale and the dividend
‘on the C-scale— you will always find the quotient on
the same scale as the dividend! You must remember
that.
Reciprocals
Every time you carry through a division on the C-
and D-scales you are actually performing two divisions,
3one of which is upside-down, so to speak, compared ta
the other.
Let us try 23 + 5. We will solve it first in the ordinary
way. Since 5 is the divisor, we place C-5 over D-23 and
under C-1(right) we find D-46. Keeping the order
magnitude in mind, we see that 23 +5 = 4.6 (Figura
53).
Take another look at the slide rule in the setting o
Figure 53. That is exactly the setting we would want if
we took the example 5-23 and did it upside-down
placing the divisor, 5, on the D-scale directly under #
dividend, 23, on the C-scale, In that case, we will fing
the quotient on the C-scale, where the dividend is, and
over D-1, we will find a reading just to the left of C-214
If we call it C-2175 and adjust the order of magnitude,
we decide that 5 + 28 = 0.2175.
The same setting on the slide rule that gives us the
quotient of any division, say a+ b on the D-scale, will
give us the quotient of the division b + a on the C-scale.
We can write divisions as fractions. We can write
a+b as a/b and b +a as b/a, We can therefore say
that the same setting of the slide rule that gives us the
value of the fraction a/b on the D-scale will give us the
value of the fraction b/a on the C-scale,
‘The fraction b/a is called the reciprocal of a/b and
vice versa. It is often important in computations to
determine reciprocals, and here is one method for doing
that on the slide rule.The most important reciprocals are those which in-
volve fractions with a numerator equal to 1. The recip-
rocal of % is %, but it is customary to drop a 1 that
appears in the denominator and treat such a number as
consisting of the numerator alone. In other words, % is
usually written simply as 4 and we can say that the recip-
rocal of 4 is 4. In the same way, the reciprocal of %x is
18, that of %» is 29 and so on. On the other hand, the
reciprocal of 17 is 47, that of 34 is %44 and so on.
Reciprocals expressed in fractional form are no prob-
lem, but what if you wish them converted to decimals.
The reciprocal of 29 is 4v, but what is the decimal form
of 0?
Suppose we treat 4o as a division, as 1 + 29, Since
29 is the divisor, place C-29 over D-1 and under C-1-
(right) you will find D-345 (Figure 54). Keeping the
order of magnitude in mind, you decide that 0 =
0.0345. If you had placed C-29 over D-I (right), you
would have found, under C-1, D-345, and have obtained
the same answer.
Looking at it in general, we can say that if we place}
€-x over D-1, we will find D-1/x under C-1(right). If]
116
we place C-r over D-1 (right), we will find D-1/z under
CL
And, of course, if 1/x is the reciprocal of x, then we
can also say that x is the reciprocal of 1/x. If Y=
0.0845, then 1/0.0845 = 29.
Proportion
It is possible to combine multiplication and division.
Suppose you were faced with the problem: 13 x 5.4
0.83.
Place C-1 over D-13 and move the hairline to C-54.
The hairline now marks off a D-reading which is the
product of 18 and 5.4, a product which you needn't look
at, but which now becomes the dividend for which 0.83
is the divisor. Leaving the hairline where it is, you move
C-83 directly under the hairline and under C-1 (right)
is D-846 (Figure 55).* Now it is necessary to place the
decimal point. Consider that 13 X 5.4 + 0.83 might be
approximated as 10 X 5 +1, which comes to 50. Ac-
‘IF the problem had been 13 x 5.4 + 0.89, C-L(right) then the
D-reading under it would have been obscured by the glass edge
‘of the indicator assembly —at least on my slide rule. Don't
panic, however, Just move the indicutor assembly and look!
1cepting that as the order of magnitude, you decide ¢]
18 X 5.4 + 0,83 = 846.
Some problems involving a combination of multiplit
cation and division form a special case.
Consider, for instance, two fractions, a/b and c/d]
which are equal in value. We can then say that a/b
¢/d. We can refer to such an equality of fractions as
proportion, and read it “a is to b as c is to d.”
‘Tt may happen that you know the value of three of
four parts of a proportion and may want to find th
value of the fourth. If one of the numerators is
the equation becomes a/b =2/d. Solving for x by
ordinary algebraic methods, we find that x = ad/b or
xd+b. If one of the Senonaioators is unknown, w
have a/b = c/x which means that x = be/a or b X ¢ +
a. In either case, we can obtain the value of x on the
slide rule in the manner described just above for com-
bining multiplication and division.
However, problems involving proportion lend them-
selves to particularly easy treatment on the slide rule.
Let’s substitute numbers for the three known quanti.
ties and take % = 9/x as an example. Consider the frac-
tion % first. We know that its decimal equivalent is 0.4
and if we wish we can check that on the slide rule by
considering the fraction to be 2 5. We place C-5 over
D.2, and under C-1(right) we find D-4,
But this same- quotient would be given by any other
fraction that was equal ito 0: For instance to = 0.4,ivi ah al
Figure 55
the = 0.4, 04, %o=044, and so on, If, then, we
tried to divide 4 by 10, or 12 by 30, or 1 by 2.5, or 6 by
15, we would end with the same setting of the slide rulq
with C-L(right) over D-4. Having achieved this setting
in the case of one such fraction, say %, we ought to find,
on that same setting, all the other divisions I've men.
tioned and, indeed, any division at all that yields 0.4
the quotient. Sure enough, if you'll look at Figure 56,
you will find that just as C-5 is over D-2, you have C-1
over D-4, C-8 over D-12, C-25 over D-1, situations whicl
are equivalent to %, Yio, "%o, and %s.
To be sure, we can't foals on the C- and D-scaleg
pubes
because C-15 is far to the left of the end of the D-scale.
However, if we switch to the CF- and DF-scales (see
page 99), we will find CF-15 under DF-6.
In short, once you set up a given fraction on the slide
tule, with the denominator on the C-scale directly over
the numerator on the D-scale, then all other fractions of
equal value are similarly given (denominator on C-scale,
numerator on D-scale) on other parts of the slide rule.
If, on the other hand, you wish to place the fraction with
the numerator on the C-scale over the denominator on
the D-scale, then you will find other fractions with
numerator and denominator simitarly placed.
1rdusyosua
nuvi
In a proportion, two fractions are set equal to each|
other. If one fraction is set upon the C- and D-seales in|
the manner described, the other can also be found.
If we go back to %=3/x and place C-5 over D-2,
then we should expect to find C-x over D-3. We move’
the hairline to D-3 and immediately above it is C-75.
Since 2 and 5 in the first fraction are of the same order|
of magnitude, 3 and x should also be of the same order}
of magnitude. We therefore place the decimal point so|
as to say % = 8/75.
In the same way, if we are faced with the proportion
5.12/27 = 1/343, we place C-277 over D-5.12 and move
the hairline to C-343, finding that to be over D-633 (see
Figure 57). Since 5.12 in the first fraction is two orders
of magnitude smaller than 277, we expect x to be two
orders of magnitude smaller than $4.3. We decide, then,
that 5.12/277 = 0.633/34.3.
Inverse Scales
The key respect in which slide rule division differs
from slide rule multiplication is that in multiplication,
we set C-1 and find the answer under the multiplier,
while in division, we set the divisor and find the answer
under C-1.
This is to be expected, since division is the inverse of
multiplication, When I was explaining the workings of
the addition rule, there was this same inverse effect in
em
Figure 57
122
wpa eynanan
vont Rabat tostbtig Bich den
etn i fief hanna nab
tpabrgondaad
jefe
the mirror image of the CF-scale, in fact, and is, there-
fore the CIF-scale (Figure 58.)
‘To distinguish these inverse scales from ordinary ones,
and to make sure that the casual slide rule user notices
something is different, the Cl- and CIF-scales are usually
printed in red, whereas all the other scales I have dis-
cussed are printed in black.
The CF-scale can be used in division, just as an
verted scale was used in subtraction on the addition
rule (see page 37).
Consider the problem 12+ 4. If this is carried
through in the ordinary manner, using the C- and D-
scale, C-4 is brought over D-12 and under C-1 (right)
is found D.8, showing that 12 + 4 = 8.
Suppose, however, we use the Cl-scale along with the
D-scale. Again, we will deal with the problem 12 + 4.
This time we place the CI-1 over D-12 and move the
connection with subtraction as compared to addition
(see page 33).
Once you are thoroughly used to working the slide
rule in multiplication and division, you will find .no
difficulty in setting C-1 in multiplication but not in divi-
sion. Nevertheless, there is some usefulness to preparing
a situation in which you can set C-1 for both division
and multiplication.
To make this possible, we must do for division what
we did for subtraction in the addition rule. We must
prepare an incerse scale, one that “runs backward.” An
inverse logarithmic scale would have its numbers crowd-
ing to the left, in mirror-image to a normal scale.
Such an inverse scale is placed in my slide rule imme-
diately above the C-scale. It is a “C-inverse” or Cl-
scale. Above the Cl-scale and immediately under the
CF-scale is another inverted scale, a folded one. It is
Figure 59hairline to Cl-4. Under it is D-3, the quotient (Figure
59). Suppose we had tried to solve 45 + 9 in this way.
We place CI-1(right) over D-45, but find we canhotf
move the hairline to CI-9, for that is far off the D-scale.
We switch to the CIF-scale, therefore. Placing the hair-|
line on CIF-9, we find it under DF-5, so that 45 + 9 = 5|
(Figure 60). (Had we placed CI-1 over D-45, we would
not have had to switch to the folded scales.)
By using the inverse scales, in other words, we evolved
a routine in which we set the end of a scale (C-1 in
multiplication and CI-1 in division) in carrying throug
both multiplication and division.
The inverse scales can be used for another purpose,
too. To see what that is, let us consider some reading
on the C-scale, which we can call x, and the reading
immediately above it on the Cl-scale, which we will
call y.
The distance of x from C-1 on the left represents the
logarithm of x. The distance of y from CI-1(right) —
remember the Cl-scale is a “backward” one — represents
the logarithm of y. Therefore, the entire stretch from
C-1 to Cl-1 (right) is equal to log x + log y (see Figure
61).
But the distance represented by the entire stretch
from C-1 to CI-1(right) is equal to log 10. We can
therefore say that log x + log y = log 10. If we switchfrom logarithms to antilogarithms we say that xX y=
10, or x= 10/y.
We can, of course, adjust the decimal point to suit]
ourselves, For instance, we can move the decimal point}
of y in such a way as to divide its value by ten. In that}
case, in order to keep the value of 10,’y unchanged, we|
must also divide the numerator of the fraction by ten]
We end by saying that x = I/y. (For that matter we cat
also, with full justification, say that x= 100/y or x
0.001/y, but it is most convenient to say that x = 1/y.
If x= 1/y, then x is the reciprocal of y and vice versa,
Tt follows then that when the hairline crosses the slide,|
it marks off simultaneous readings on the C-scale and|
Cl-scale that are reciprocals of each other. You can see}
for instance that CI-2 is directly over C-5. This tells us|
(if we make sure to keep the decimal point in the proper|
position) that the reciprocal of 2 is 0.5 and vice versa,
since 2= 1,05 and 0.5 = 14. The same reading tells us}
that the reciprocal of 20 is 0.05, that the reciprocal off
5 is 0.2, that the reciprocal of 0.005 is 200, and so on,
since Yo = 0.05, % = 0.2, and 1/0.005 = 200.
For none of these reciprocals is the slide rule reall
128
needed; we can work it out in our heads without trouble.
The principle is established, however.
Suppose, now, you wanted the reciprocal of 485. First
let's place the decimal point. Fractionally, the reciprocal
of 485 is Mss which is slightly larger than ‘oo. Since
» = 0,002, we expect the reciprocal of 485 to be
slightly larger than 0.002.
Now place the hairline on C-485 and you will find it
to be marking CI-206 at the same time (Figure 62).
The reciprocal of 485 is therefore 0.00206. (If you had
placed the hairline on CI-485, you would have found it
to be marking off C-206. The same is true of the CF-
scale and the CIF-scale taken in combination. A reading
of CF-485 corresponds to one of CIF-206, and a reading
of CF-206 corresponds to one of CIF-485.)
Reciprocals can also be found on the C- and D-scales
by the method I described earlier (see page 130). The
use of the inverse scales, however, is the simpler for such
a purpose.
I will describe one other adventure in reciprocals. As
you know, for any C-reading, say C-x, the corresponding
CF-reading is xx (see page 101). In other words, cor-
129responding to C-2 is CF-628, so that 2r. = 6.28.
Since the CIF-readings are the reciprocals of the cor-
responding CF-readings, for the setting C-x the corre-
sponding reading on the CIF-scale is equivalent tod .
Let's take a specific example, Corresponding to C-2,
we have CIF-1591 (Figure 63). This means that =
0.1591. Placing the decimal point is not difficult, if we
remember that x is close to 8. In that case - is approxi-
mately % or 0,167. That sets the order of magnitude at
once.
Similarly, if you want the value of ae place the
hairline of C-82 and find the simultaneous reading CIF-
888, Since 8.22 is close to 25, we can say that 5:
Powers
The Other Side
So ran, I have described seven scales, one under the
other, on the slide rule. On the upper part of the body
is the DF-scale. On the slide are the CF-scale, CIF-
scale, Cl-scale, and C-scale. On the lower part of the
body are the D-scale and the L-scale,
There is room for more and, indeed, one slide rule I
use has several scales in addition to those 1 have listed.
These additional scales, however, involve operations
which will not be taken up in this book, and therefore,
these scales will not be discussed.
Even so there remains the other side of the slide rule
which is available for other scales, some of which can
be used for operations we will consider in this book.
Let’s take up the other side of the slide rule, then. We
can begin by pointing out that there are two ways of
reversing the slide rule. It can be rotated horizontally
in a long swing, or vertically in a short flip.
Obviously, the short fip is easier and it is the method
of rotation used. The proof of that is, that if you rotate
a slide rule horizontally, you will find that the markings
on the reverse side are upside-down. If you flip it
vertically, the markings on the reverse side are right-
side up. You are clearly expected to perform the latter
movement.
131In flipping the slide rule, however, we reverse the
Position of the top and bottom of the body. What was
the upper body in front becomes the lower body in back
and vice versa. You can check this by placing one
finger on the upper body and keeping it there while you
flip.
The scales on the other side of my slide rule are, with
one exception, different from those on the front. That
‘one exception is the D-scale, which is present on the
reverse side in the same D-position (see page '67) as
in front.
To distinguish this D-scale from the one we have been
using previously, we will call it the D(back)-scale.
Notice that the D(back)-scale is not on the same part
of the body as the D-scale. They both seem to be on
the lower body if we look at each separately, but in
order to look first at one, then at the other, we must flip
the slide rule. If you place your thumb on the D-scale
and your forefinger on the D(back)-scale, you will see
that the two are on different parts of the body.
If the slide rule is properly adjusted; if one half of the
body is fastened directly over the other and if the two
glass windows of the indicator assembly are properly
positioned, so that the hairline of one is directly opposite
the hairline of the other, then the two D-scales will be
coordinated. If the hairline gives a reading of D-x, it
will be found to give a reading of D(back)-x, when the
slide rule is fipped.
‘The D(back)-scale acts as a connecting link between
the two sides of the slide rule. Most often, the answer
obtained in an ordinary multiplication or a division will
be found on the D-scale. If there is then anything you
wish to do to that answer which will involve scales
found only on the other side, you need only fip your
slide rule. There is the answer found on the D-scale,
properly marked off on the D(back)-scale, without any
necessity on your part to make a new setting. And you
may then continue.
Under the D(back)-scale, for instance (on my slide
tule), is a Di-scale, which is the red-marked inverse of
the D-scale, just as the Cl-scale is the red-marked inverse
of the C-seale (see page 144).
‘The D(back)- and DI-scales can be used, together, to
find reciprocals, just as the C- and Cl-scales or the CF-
and CIF-scales can be so used (see page 128).
‘There are two advantages, though, to using the D-and DI-scales. In the first place, if you are working on
the other side of the slide rule, you can find your recip-
rocals right there.
‘The less obvious one (for my slide rule) is that the
Dl-scale is the only inverse scale on the body and not
on the slide. This means that reciprocals can be found
regardless of the position of the slide,
Suppose, for instance, you wanted the solution to a
problem like go 23.- To obtain the product of 6.2 X
18 is simple, but once that is obtained what you want is
the reciprocal.
enttet
supp ree
eoeapeomymerngearpenpmgeogg
camming eeeLet us begin. We place C-1(right) over D-62, move fl
the hairline to C-18 and find that to be over D-1116.
What we need, now, is the reciprocal of 1116. This
means, if we are to use the C- and CI- scales, or the CF-
and CIF-scales, that we must move the hairline to 1116
on one of those four scales.
Yet that is not necessary. If we flip the slide rule, the
D-reading is duplicated in the D(back)-reading and the
DI-reading is the reciprocal of that. We need move
neither slide nor hairline.
- 1 4
In shor, let's try to solve the problem g53-y-g again.
Place C-1(right) over D-62, move the hairline to C-18,
flip the slide rule and note that the hairline marks DI-
896 (Figure 64). Next we will check the decimal point.
Since ga Ayg can be approximated as gig, the
answer must be about 42 or 0.083. Hence, eee =
0.0896.
The A- and B-scales
On the reverse of the slide rule in the A- and B-posi-
tions (see page 66) are two scales actually named for
those positions, the A-scale and the B-scale ( Figure 65).
The A- and B-scales are spew variety. Except for the
ae ny
Donte
m0 Twn $ tnt
L-scale, all the scales 1 have discussed so far have really
been members of a single family represented most
clearly by the C-, D-, and D(back)-scales. The CF-
and DF-scales differ from these only in that they start
at the middle; the CI- and Di-scales only in that they
run backward; the CIF-scale only in that it starts at
the middle and runs backward also. All these, however,
have the same system of primaries, secondaries, and
tertiaries.
Not so the A- and B-scales. They begin at the left at
1 and pass through the units, reaching 1 again in the
middle. There is then room for an exact repetition,
reaching 1 still again at the right end. It is as though
the C- and D-scales have been compressed to half their
length so that two of them will fit into the usual slide
rule length.
We can refer to the two identical halves of these
scales, as the A(Jeft)-scale and A(right)-scale, and, of
course, the B(Jeft)-scale and B(right)-scale.
Let's stop to consider what this means. You may
remember that I described the construction of the C-
and D-scales by lining up their markings as antiloga-
rithms matching the logarithmic readings of the L-scale
(see page 67).
The L-scale runs from 0.0 to 1.0, but what if we were
to construct such a sole swith the units spaced moreclosely together, so that there would be room for aj
stretch from 0.0 to 2.0. If antilogarithms were lined up.
with such a “double L-scale,” the results, as shown in|
Figure 66, would be a logarithmic scale running from!
Ito 100.
Since the slide rule takes no account of the decimal
point, the primaries of the A- and B-scales are not:
marked 1, 2, 8, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, 50,
60, 70, 80, 90, 100, but 1, 2 8 4, 5, 6 7, 8, 9,
1, 2, 8, 4, 5, 6.7, 8, 9, 1. There is thus an A-I, an A-l-
(middle), and an A-1(right.)
On the whole it is best to let the A(left)-scale repre-
sent the stretch of the numbers from 1 to 10 and the
A(right)-scale from 10 to 100, varying the decimal
point two places at a time to get other stretches, Thus,
the A(left)-seale includes not only, let us say, the num-
ber 5, but also 500; 50,000; 5,000,000; and so on. Work-
ing in the other direction, it also stands for 0.05, 0.0005,
0.000005, and so on. On the other hand, the A(right)-|
scale, includes 50; 5,000; 500,000; 50,000,000; and so}
on; as well as 0.5, 0.005, 0.00005, and so on,
This is not really difficult to remember. You will
notice that the A(left)-scale includes the numbers with
an even number of zeros neighboring the decimal point!
when written in the’ form given above. Thus, 500 has
two zeros neighboring the decimal point, as has 0.05;
and both 50,000 and 0,0005 have four zeros neighboring}
the decimal point. (Naturally, numbers less than one|
have to be written with the preliminary zero, as 0.05)
and not as .05 if this rule is to hold.) The number 5 has|
no zeros and if this is considered an even number
zeros, this, too, fits the rule.
On the other hand, the A(right)-scale includes the}
numbers with an odd number of zeros neighboring the
decimal point, Thus, 50 and 0.5, both have one zero
neighboring the decimal point, 5000 and 0.005 both
have three, and so on. Alll this, of course, holds for the
B-scale as well.
‘The A(left)-scale crowds the full stretch of values
found in the C- and D-scales into half the usual length.
‘This means that the A(left)-scale cannot he subdivided
as finely as the C- or D-scale (or any of the others of
that family). There isn’t the room.
The space between the numbered primaries of the
A(left)-scale are divided into secondaries representing
tenths, which are spaced more closely, of course, than
are the secondaries in the C- and D-scales. Even be-
tween the primaries 1 and 2, where there is the most
room, the secondaries are crowded too closely together
to allow a convenient marking off by number as is done
in the C- and D-scales and the rest of that family. None
of the secondaries of the A- and B-scales are numbered.
The space between the secondaries lying in the range
from 1 to 2 on the A(left)-scale are split into five sub-
divisions marked off by tertiaries representing 0.02 each.
From 2 to 5, the space between the secondaries are
divided in half, with a single tertiary representing 0.05.
Between 5 and 1, there are no tertiaries at all. The
markings on the A(right)-scale and of course on the
B(left)- and B(right)-scales are identical with those
on the A (left)-scale.
If you make use of the A- and B-scales, it is important
to keep a wary eye on their markings and make sure
you do not carry over the habits developed on the C-
and D-scales.
You can multiply and divide on the A- and B-scalesFigure 67
sa
precisely as you can on the C- and D-scales. There are
both advantages and disadvantages to this, I will
mention the advantages first.
Since there are two scales, a left and a right, included
in the A- and B-scales, there is no likelihood of the slide
moving off the scale and you do not have to interchange
indices.
Suppose, for instance, you multiply 6 by 2 on the C-
and D-scales and carelessly begin by putting C-1 over
D-6. You must next move the hairline to C-2, which,
however, is well beyond the right edge of the slide rule.
Nor can you switch to the CF- and DF-scales where, in
this case, CF-2 is even farther beyond the right edge.
There is no recourse but to switch indices.
If you did the same thing on the A- and B-scales, you
would have no difficulty. It is the B-scales that is on
the slide, so it is that which you move. Place B(left)-1
under A(left)-6 and now move the indicator to B(left)-2.
To be sure, B(left)-2 is beyond the right end of the
‘A(left)-scale, but here we have an A(right)-seale to ff
take up the slack (Figure 67). B(left)-2 is under A-
(right)-12 so 6 X 2 = 12. It is for this reason that there
no need for a folded A- or B-scale,
The A- and B-scales can be used for division, too, in a}
manner analogous to the C- and D-scales. Here, too,
the divisor is best placed on the slide; that is, on the
B-scale.
Consider the problem 5.4~+ 2.2. The divisor is 2.2
so you place it on the B-scale; B(left)-22 is put directly
under A(eft)-54 with the help of the hairline, and over
B(left)-1 is A(left)-245 (Figure 68). Adjusting the
decimal point, you decide that 5.4 + 2.2 = 2.45.
The A- and B-scales can also be used to find recipro-
cals in just the same way that the C- and D-scales can
be used (see page 113).
Squares
The A- and B-scales can do everything, it seems, that
the C- and D-scales can do, without the danger of having
to switch indices in mid-problem. Nevertheless, there
is a serious disadvantage to their use.
Since the markings on the A- and B-scales are only
half as far apart as the equivalent markings on the C-
and D-scales, the A- and B-scales cannot be read as
closely. This means that any answers that you get on
the A- and B-scales are not as accurate as those you
get on the C- and D-scales.
“1
silica EENsomone igen
mens ona
This question of accuracy is so important that ordinary
multiplication and division should always be done on
the C- and D-scales and never on the A- and B-scales.
Increasing the accuracy is so desirable that one should
willingly pay the price of an occasional shift of indices,
or of having to transfer attention to the folded scales.
Yet I am not suggesting that the A- and B-scales be
completely ignored, that they never be referred to. The
A-scale, at least, has a particular convenience that is
quite sufficient to justify its existence.
To see what that is, let me remind you again that the
C- and D- scales are built up on the L-scale, which runs
the length of the slide rule with a range from 0.0 to 1.0,
while the A- and B-scales are built up on a similar
“double scale” with a range from 0.0 to 2.0 (see page
187).
Figure 68
Since the double scale covers twice the range in the
same distance, it moves, so to speak, twice as fast. Ata
given point where the L-reading is, let us say a, the
“double L’”-reading would be 2a.
Let us now suppose that the hairline is over a partic-
ular D(back)-reading and is simultaneously over a
particular A-reading; we can call them D(back)-x and
A-y. The distance of D(back)-» from D(back)-1 rep-
resents the logarithm of x and the distance of A-y from
A(left)-1 represents the logarithm of y on the double
scale. It is the same distance in both cases, but the
double scale logarithm is twice the size of the ordinary
logarithm at the same point, so that we can say log
y=2 log x, or log y = log x + log x (Figure 69).
If we shift to antilogarithms, remembering to convert
an addition into a multiplication, we have y = x X x, or
|to use the usual mathematical terminology, y = x*.
What we have now shown is that when the hairline
marks off simultaneous readings on the D(back)-scale
and the A-scale, the A-reading is the square of the
D(back)-reading.
To determine a square, then, we need only adjust
the hairline without moving the slide. To determine the
square of 5.1, place the hairline over D(back)-51 and
note that it also gives a reading of A(right)-26. We
can establish the order of magnitude by remembering
that 5.1 is not much different from 5 and that the
square of 5 is 25. Therefore 5.1? = 26, .
In the same way, D(back)-884 corresponds to
A(right)-147 (Figure 70). To place the decimal point,
we substitute 4 for 3.84. Since we know that 4? = 16,
that 40? = 1600 and that 0.4? = 0.16, we can tell that
3. 4.7, that 88.4° = 1470, and that 0.884? = 0.147.
It is helpful, in connection with finding squares and
placing the decimal point, to remember the odd-even
rule concerning the right and left halves of the A-scale.
The A(right)-scale involves numbers with an odd num-
ber of zeros neighboring the decimal point. For that
reason, you may expect 7 jzeedings to be of the order
Teen
a=
ots
of magnitude, let us say, of 10 or 1000, but never 100.
Therefore since the squares of 384 and of 51 fall in the
A(right) range, their squares might be 14.7 or 1470 in
H the first case or 26 or 2600 in the second, Never, no
matter how you place the decimal point in $84 or 51,
can the square turn out to be 1.47 or 147 or 2.6 or 260.
Another point to remember is that the D(back)-scale
is not an absolute necessity for determining squares on
the A-scale. The D-scale itself can also be used. Once
a D-reading is set, one need only flip the slide rule to
the other side to find the squares marked off as a
simultaneous A-reading.
To be sure, squares can be determined without the
A-scale too. If 8.84 is considered to be 3.84 x 3.84,
that multiplication can be carried through on the C-
) and D-scales in the ordinary way. If this is done, the
answer is obtained with greater accuracy, because the
final reading is made on the D-scale and not on the
less finely divided A-scale.
If, indeed, we carry through 3.84 X 3.84 on the C-
and D-scales by putting C-1(right) over D-384, then
moving the hairline to C-384, we will find, under it,
| D-1475. We will conclude, that 8.84? = 14.75. This is
ee eecloser to the truth than the 8.84? = 14.7 determined b
use of the D- and A-scales; for the true answer, worl
out in full, is 14.7458.
Nevertheless, working out the square by use of
A-scale involves no motion of the slide, and for quicl
work that maneuver comes in handy.
By substituting the Dl-scale for the D(back)-scale
we can get the square of a reciprocal: say (By-
can set the hairline on DI-48 and obtain a simultaneo
reading of D(back)-208, which gives us the recipr
& However, we needn't stop to work out the deci
place or even look at the digit combination. We pass
directly on to the A-scale, which will give us the square|
of the D(back)-reading and hence the square of the’
reciprocal of the DI-reading.
We find the A-reading to be A(left)-433 (Figure 71),
and now we are ready for the decimal point. In place of.
48, let us take 5. We therefore have (4)? which equals
(0.2)? or 0.04, an order of magnitude with two zeros
neighboring the decimal point (checking the fact that
we ended with an A(left)-reading, which requires an!
even number of zeros), We conclude then that (745
z
= 0.0488.
The fact that the slide rule gives us a method for
quickly determining squares means that we can carry
through a multiplication or a division and find the
square of the result at once.
If there were a C( back )-scale adjoining the D( back )-
scale on the reverse side of the slide rule (as there is, in
some slide rule designs), we could run the multiplica-
tion or division on those scales, get the product or
quotient on the D(back)-scale and take the simultaneous
A-reading,
On my slide rule, however, there is no C( back )-scale.
The reverse of the slide, where such a scale should be,
is given over, instead, to several scales involving trigo-
nometrie functions which I will not discuss in this book.
Fortunately, this introduces no great difficulty. We
can use the C- and D-scales on the face of the slide rule
instead.
Suppose, for instance, we want to solve the problem
(9.44 X 2.75)®. Beginning at the front of the slide rule,
we place C-I(right) over D-944, move the hairline to
C-275, and find a D-reading under it that represents the
product of 9.44 X 2.75. However, we are not interested
in the product but in its square, so we needn't even
147Figure 72
bother looking at the D-reading, Instead, we flip the |
slide rule in search of the A-reading and find it to be
A(left)-673.
We get the order of magnitude by substituting 10 for
9.44 and 8 for 2.75. The problem becomes (10 X 8)?
= 80? = 900, which checks with the fact that an A(left)-
reading requires an even number of zeros in the
neighborhood of the decimal point. We conclude, then,
that (9.44 X 2.75)? = 678.
‘You can use the same principle to solve (6.3 + 1.18)?,
provided you remember to manipulate the slide rule in
such a fashion as to let the quotient of 6.3 + 1.18 fall
upon the D-scale. Only if it falls upon the D-scale can
the square of the quotient simultaneously be found on
the A-scale. However, if we follow the rule of placing
the divisor on the C-scale (see page 128), this point will
take care of itself.
Since 1.18 is the divisor, it goes on the C-scale. Place
C-118 over D-63, Move the hairline to C-1 and under it
is D-535, which represents the quotient. Without
bothering with the D-reading at all, however, you flip
the slide rule and find A(right)-285. To place the
decimal point you replace 6.3 by 6 and 1.18 by 1. The
problem becomes (6+ 1)?=6=S6. This order of
somali ttt Le; hich checks with the feot thet
the final reading was found on the A(right)-scale,
requiring an odd number of zeros in the neighborhood
of the decimal point. Therefore (6.3 -- 1.18)? = 28.5.
Cubes
An exponential number like 2* is said to be a power of
x, The higher the value of the exponent, the higher the
power. The average slide rule can reach a higher power
than the square; it can, in fact, handle the cube (x*) just
as easily.
The square involved a double L-scale so it is not
unexpected that the cube will involve a “triple L-scale.”
Since the ordinary L-scale covers the range of logarithms
from 0.0 to LO, the triple L-scale will cover the range
from 0.0 to 8.0. In terms of antilogarithms this is the
range from 1 to 1000.
The antilogarithm scale with markings from 1 to 1000
is present on the front of my slide rule, all the way at
the top, and is the K-scale (Figure 72). (In another
slide rule I have, it is on the hack, all the way on the
bottom.) If you follow its markings you will see that its
primaries represent 1, 2, 8, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40,
50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800,
900, 1000. Since the decimal point is arranged to suit
“9i
ourselves, it is only necessary to write the digits: 1, 2, 8, 4,
5, 6, 7, 8, 9, 1, 2 8, 4, 5, 6, 7, 8,9, 1, 2,8, 4,5, 6,7, 8,9,
L
Just as the A- and B-scales are divided into two iden-
tical halves, the K-scale is divided into identical thirds,
so that there is a K(left)-scale, a K( middle )-scale, and a
K(right)-scale.
The decimal point jumps three places at a time in
each of the three sections, thus:
K(left)-scale
K(middle)-scale
K(right)-scale
0.002 2 2,000 2,000,000
0.02 20 20,000 20,000,000
0.2 — 200 200,000 200,000,000 etc.
The individual thirds of the K-scale are even more
crowded than the halves of the A- and B-scales. The
spaces between 1 and 3 are marked off by secondaries
with values of 0.1 and tertiaries with values of 0.05. The
spaces between 3 and 6 are marked off by secondaries}
only, with values of 0.1 each; and the spaces between 6
and 1 by secondaries with values of 0.2.
The K-scale exists by itself; there is no similar scale
that can be moved along it as the B-scale can move along]
the A-scale, or the C-scale along the D-scale. For that
reason there is no way of using the K-scale directly for
ordinary multiplications and divisions, and that is just
as well, for the K-scale would be extremely inaccurate
for that purpose.
The important use for the K-scale is in connection
with cubes. In fact, the “K” of the K-scale stands for
Kubus, the German word for cube.
The manner in which it was shown that the readings
150
on the A-scale represent the squares of the correspond-
ing readings of the D(back)-scale (see page 144) can
be used (with appropriate modification for a triple L-
scale, rather than a double) to show that K-readings
represent the cubes of corresponding D(back)-readings.
For instance D-2 corresponds to K(left)-8; D-43
corresponds to K(middle)-8 and D-928 corresponds to
K(right)-8. Remembering the discussion on decimal
points above, we see that 2” = 8, 4.8* = 80, and 9.28" =
800.
The method for solving problems such as (64 X 0.4)*
and (3.11 + 2.6)? is just the same as for solving anal-
ogous problems involving squares (see page 144). We
merely look for the K-reading, rather than the A-reading.Roots
Square Roots
Wuen we dealt with addition, we found we could,
also deal with its reverse — subtraction. Again, in deal-
ing with multiplication, we dealt with its reverse —
division. Now, in connection with powers, we find that
the reverse operation can also be dealt with, and it in-
volves roots.
The square root of a is that number which, when
multiplied by itself, yields a as a product. In other!
words if bX b=a (an equation which can also be
written b? = a), then b is the square root of a, and this
can be written b = Va, The square and the square root
are thus opposite sides of a coin, so to speak. If a is the
square of b, then b is the square root of a.
We can make this plainer with numbers, perhaps.
Since 5 X 5 = 25, 25 is the square of 5, and 5is the square
root of 25, Since 6 X 6 = 36, then 36 is the square of 6,
and 6 is the square root of 36.
To find the square of a number, using the C- and D-
scale, as I pointed out earlier (page 144), is simple
enough. To determine 7.8%, you need only solve for
7.8 X 7.8 in the usual way and find the answer to be 60.8;
To find the square root of a number using the C- ani
D-seale is quite another thing. Suppose you want the
solution to V7.8. That means you want a number whic
when multiplied by itself, will yield 7.8 as a -produ
You find that product onthe D-scale and mark D-
with the hairline.
Next you must adjust the slide in such a way that the
C-reading at the hairline over D-78 must be the same as
the D-reading under C-1. By careful manipulation of
the slide you will find that when C-279 is over D-78, C-1
is over D-279. That means that 2.79 X 2,79 =7.8 and,
therefore, that V7.8 = 2.79,
Adjusting the slide to find a square root is not an.
impossible procedure, but it is a tedious one, and must be
carried through carefully, with frequent glances back
and forth between C-1 and the hairline. It is not at all
the usual flick-Aick of the slide rule as in carrying through
so many other computations,
Consequently, if another technique will suffice to
find the square root, it should be used.
‘The answer lies with the A-scale and with the knowl-
edge that finding a root is the reverse of the operation
of finding a power. Since the A-reading is the square of
the simultaneous D(back)-reading, the D(back)-read-
ing is the square root of the simultaneous A-reading.
To find the square root of 7.8, then, it is only necessary
to place the indicator on A(left)-78. The simultaneous
reading of D(back)-279 (Figure 74) is sufficient to tell
us that V78=2.79, with that single setting of the
indicator, Not only is it not necessary to adjust the
slide carefully back and forth to find the square root, it
isn’t necessary to move the slide at all.
By combining the A- and Dl-scales, we can obtain
the reciprocal of the square root. if we wanted =, we
would set the hairline at A(left)-78 and note the simul-
taneous reading of DI-858 (Figure 73). We therefore
conclude that a =0.858, (To locate the decimal
153ie
vn
point, we need only replace 78 by the fairly close ap-
1.
a — = % = 0.838.
proximation 9. We then see that 7 = % = 0.998.)
Notice that we determined the square root of 7.8 to
be 2.79 by setting the hairline at A(left)-78. ‘What if we
had set the hairline at A(right)-78? If we do so we
would find a simultaneous reading of D-878. The dif-
ference is a matter of decimal point. We are now deal-
ing not with 7.8 but with 78, and we can say that
V78 = 8.78.
In finding a square root, we must first determine
whether to use the A(left)-scale or the A(right)-scale.
To do that we deal only with the first digit of the number
whose square root we want, converting the other digits
into zero, We then count the number of zeros in the
neighborhood of the decimal point, and for even num-
bers use the A(left)-scale, and for odd numbers the
‘A(right)-scale, in accordance with the rule given on
page 161,
For the square root of 723, we use the number 700 as
a guide. Since there are an even number of zeros, we
use the A(left)-scale. For the square root of 0.562, we
use 0.5 as the guide. Now we have an odd number of
zeros and use the A(right)-scale.
If your memory fails you, you can always use
154
method of approximation, Suppose that you want the
square root of 67. If you use the A(left)-67 setting,
you obtain a reading of D(back)-259. On the other
hand, if you use the A(right)-67 setting, you obtain a
reading of D(back)-819. You need not hesitate between
the two. Since 67 is not far removed from 64, and you
know that the square root of 64 is 8 (since 8 X 8 = 64),
you have no need to pause in saying that V67 = 8.19.
(On the other hand, V6.7 = 2.59.) Of course, if you
had treated 67 as 60 and noted the odd number of
zeros — just one, that is— in the neighborhood of the
decimal point, you would have used the A(right)-scale
at once.
Cube Roots
Just as we could reverse the procedure of finding
squares in order to make it possible for us to find square
roots, so we can reverse the procedure of finding cubes
so as to make it possible for us to find cube roots.
The cube root of b is the number which, when multi-
plied by itself twice, gives b as a product. IifaxaXx
a=b (which can also be written a° = b), then a is the
cube root of , or, as it is usually written, a= YB, Cubes
and cube roots are inverses of each other. If b is the
cube of a, then a is the cube root of b.
‘We already know that if we simultaneously take a
155,K-reading and a D-reading, the K-reading is the cube of
the D-reading (see page 176). It follows, therefore, that
the D-reading is the cube root of the K-reading.
Suppose we wanted 72. Place the hairline at
K(left)-72 and the simultaneous reading of D-193 tells
us that Y72 = 1.93 (Figure 74). The decimal point is
located by the fact that if we replace 7.2 by 8, we know
at once that the ¥8 = 2, since 2° = 8.
By combining the K- and Dl-scales, we can obtain
the reciprocal of the cube root. Since K(left)-72 gives a
1
simultaneous reading of DI-518, we know that va
0.518. Again, we locate the decimal point by replacing
72 with 8, Since 3g = ¥ = 05, we have our answer.
In finding cube roots, however, we have three K-
scales to chopse frorn for our setting. In finding the cube
root of 7.2, for instance, do we use K(left)-72, K(mid-
Figure 74
app
dle)-72, or K(right)-72? The answer lies in the table on
page 175. For numbers from 1 to 10, we use the K(left)-
scale, between 10 and 100 the K(middle)-scale, be-
tween 100 and 1000 the K(right)-scale. For numbers
from 1000 to 10,000, we are back to the K(left)-scale,
and so we progress through the three scales in order
indefinitely.
‘The same holds true for numbers less than 1. From.
1000 down to 100 we use the K(right)-scale, as I have
already said, from 100 down to 10 the K(middle)-scale,
and from 10 down to 1 the K(left)-scale. Progressing
then in this direction, we use the K(right)-scale for
numbers from 1 down to 0.1, the K(middle)-scale for
0,1 down to 0.01 and so on.
In this book 1 have dealt with twelve scales on the
slide rule. With those twelve scales, we can multiply,
divide, take proportions and reciprocals, obtain squares.square roots, cubes, cube roots, logarithms, and anti-
logarithms.
Such calculations (together with additions and sub-
tractions for which we don’t need slide rules) probably
take care of virtually all the calculations that we en-
counter in ordinary life.
‘There are, to be sure, other types of computations
that can be performed on slide rules. Those scales on
my slide rule which I haven't discussed in this book
make it possible to deal with fractional exponents, with
natural logarithms, with trigonometric functions, and
soon.
‘Then, too, special scales not found on ordinary slide
rules are adapted to the particular needs of those: who
must do calculations in special fields, such as those of
electrical engineering or of business. By bending the
slide rule into circles or spirals, a greater length of scale
(and hence greater accuracy) can be squeezed into a
particular space.
What we have described in this book, therefore, by no
means exhausts the usefulness of the slide rule.
But we have made a good beginning. Armed with the
slide rule, there will be few ordinary problems that need
hold any terrors for you. A push of the slide this way
and that, an adjustment of the hairline, and the answer
is yours!
No wonder that to many people the fee! of the ’slide
rule in the hands spells security.
As I said in the introduction — a man who must carry
out numerical computations constantly would be as lost
without his slide rule as a doctor without his stethoscope
‘or a painter without his brush,
INDEX
Asscale, 136-137
markings of, 137-139
multiplication and,
140-141
squares and, 144-145, 170
Addition, 10-11
addition rule and, 19-31
decimals and, 25-29
fractions and, 23-25
Jengths and, 11-12
Addition rule, 17-18, 1948.
construction of, 17-18
subtractions on, 32-38
Analog computer, 12
‘Answers, estimations of,
30-31
Antilogarithms, 42-47
slide rule and, 64-69
B-scale, 137-137
Body, 19
Cescale, 83-84
CF -scale, 99
Characteristic, 53-54
Cl-scale, 124-125
CIF-scale, 125
Computer, analog, 12
Cube(s), 149-151
‘Cube roots, 155-157, 184
Descale, 67
‘markings on, 68-78
D(back)-scale, 132
DF-scale, 99
Di-scale, 133-134
Decimal(s), addition of,
25-29
Decimal point, 52-57
slide rule and, 81, 88-92
Direct scale, 35
Division, exponential numbers
and, 41-42, 44
and, 44, 41-42
inverse scales and, 126-128
logarithms and, 51-52
pi (zr) and, 109-110
slide rule and, 129,
107-113
Estimations, 30-31, 90-92
Exponent, 40
Exponential numbers, 40
Folded scales, 92-99, 111
Foot-rule, 9-10
Fractions, 23-25
Hairline, 65
Index, 84
Indicator assembly, 64-65
Inverse seale, 35, 123-130
K-scale, 149
L-scale, 61-63
Left index, 84
Logarithm (3), 42, 46, 49
decimal, 49-50
decimal point and, 52-57
fractional, 48-50negative, 55-57
slide rule and, 58-64
Logarithm tables, 50-51
Logarithmic scale, 71-72
Magnitude, order of, 89-90
Mantissa, 53
Multiplication, exponential
numbers and, 40-4)
folded scales and, 102-106
logarithms and, 51
pi (mr) and, 100-101
slide rule and, 82-87
Order of magnitude, 89-90
Pi (17), 99-102
division by, 111-112
multiplication by, 100-102
reciprocals and, 129-130
Primaries, 60
Proportions, 117-123
Readings, estimated, 29-31
Reciprocals, 113-117
inverse scales and, 128-129
squares of, 146
Right index, 84
Ruler, 9
addition by, 10-11
Scale(s), 35
A, 136-137
B, 136-137
C, 83
CF, 99
Cl, 124-125
CIF, 125
D, 67
D(back), 132
DF, 99
DI, 133
folded, 92-99
inverse, 35, 123-130
logarithmic, 71-72
markings of, 60-63, 68-78
position of, 61-62, 66-67
Secondaries, 60
Slide, 19
Slide rule, 7-8, 14-16
accuracy of, 78-79
antilogarithms and, 64-68
79-80
back of, 131 ff,
cube(s) and, 149-151
cube roots and, 155-156
decimal points and, 81
division and, 107-113
logarithms and, 58-63,
79-80
ition and, 82-87
proportions and, 117-122
reciprocals and, 113-117,
128-129
square(s) and, 144-145
square roots and, 152-155
Square(s), 141-148
Square roots, 152-155
Straightedge, 9
Subtraction, 11
n rule and, 32-38
Univac, 12
You might also like Sanskrit Grammar For Beginners, in Devanagarl and Roman Letters Throughout, by F. Max Mtiller, M.a., Foreign Member of The French Instittttb, Etc. Second Edition, Revised and Accentuated. PDF
Sanskrit Grammar For Beginners, in Devanagarl and Roman Letters Throughout, by F. Max Mtiller, M.a., Foreign Member of The French Instittttb, Etc. Second Edition, Revised and Accentuated.
336 pages