Geometry in Ancient and Medieval India

34
ing enemies and so on. But all these different shapes had tO
have strictly the same area, viz. 7 Sq. purusas. Hence were
evolved methods for transforming one geometrical figure into
especially the square into other equivalent geome
another, more
figures. These constructions are given below.
trical
circle
2.10.1. To convert a square into a
exactly. What the
No geometrical method can achieve this The centre
Zulbasttras do is to give approximate constructions.
the circle is drawn
Oof the square is joined to a vertex A and
with half the side of the square combined with the excessofof
side
OA over half the side of the square,l i.e. if a is the
the square and r' the radius of the circle.
V2. a
I=
3
(241/ 2)
2 3
3.088. But
The value of calculated from this is only about sentence of
according to some of the commentators, the last anity
this rule, namely Snityã mar dalam,is to be split as Si other
and the
mandalam, when it will mean that Apastamba ap-
authors of Sulbasktras as well were aware that this was an do
understandablyV,
proximate method only. Thibaut and Bürk,
1
not accept this explanation.
2.10.2. To convert a circle into a square
wl diameter
All the three important Sulbas direct us to divide the
fi of the
into 15 parts and to take 13 of these parts as the Side side
the
square i.e., if d is the diameter of the circle and a
d of the equivalent
o
13 square,
t ls d whence=3.004

Baudhâyana gives a slightly better approximation too. qerfaafa

I.59)
(B.SI.

1 Sl. III. 2; B. SI. I. 58 and

2Ap. Sl. III. 3; B. SI. I. 60 K. SI. IIL.
and K. SI. 14
Sulbasktra Geomerry 35
(Wishing to convert a circle into a square one should divide the
diameter into 8 parts, divide one of these parts again into 29
narts and subtract 28 of these (29th parts)
sÌxth of one of these parts, the latter being together with one
eighth of that (one-sixth part) ). diminished by one
i. e., a=d (I - 28 1 1
8.29 6.8.29 + 6.8.29.8 )
This value is based on an inversion of the
and a given in connection with the problem of relation between r
How exactly the value was brought to the formcircling
of
the square.
complicated fractional expression is a matter forthis long and
but may not be of geometrical interest. speculation,
2.10.3. To convert a rectangle into asquare
rule is:

(Ap. S. II. 7)
(Wishing to turna rectangle into a square, one should
a part equal to the transverse side and the cut off
remainder should be
divided into two and juxtaposed at the two sides (of the first
segment). The bit (at the corner) should be filled in by an im
ported bit. The removal of this has been explained
already).
If ABCD is the
D, F C? rectangle,
a square AB, C, D with
side AD, is cut off from
D C it. The remaining rectangle
B, BCC, is divided into
E
two equal strips B, B, E C,
and B, BCE. The strip
B,BCE is cut off and appli
ed to the side of the square
DC, Now we get a square
of side AD, with a small
A BI Ba 3 square C, EC,Funilled
Fig. 14 up at one corner. The lar
8er Square is completed and the imported square in the corner
Temoved by the method of removing a square from asquare,
(given in 2. 11. 4.)
 Ancient and Medieval
Geometryin
Kátyáyana (IL. 2) give the
India
36 andmethod works with any rectangle
Bendhiyana(l.54)
same
Thí
Though this very long rectangle with a
the method.
Kätyayont
a
providesfor separate
bes
silra.
repeatedly hw
(K. SI. IL. 3
off
rectangle)is very long, cut formed into one big
ar he (breadth) and join the squares so
verseside remainder of the rectangle should tbe joined
the
sQuare, andthen form a square).
fits (to
tothis square as it improvement over the general method,
The method is no remainder rectangle will be equal to
side of the
since here no
square to which its strips are to be joinet
the side of the bigger
rectangle
2.10.4. To convert a square into a

(Ap. Sl. II. IP

(Wishing to convert a square into a rectangle one should make
the lateral side as long as is desired and the excess should be
joined suitably.)
We are not told how exactly the excess is to be jolleu
Thibaut and Bürk suggest that this was achieved by repeated
BCD
EC slicing and joining. IfA e of
is the square and the side
D F H the rectangle is to be AD
a rectangle A B by AtheD,re-
sliced of first. From lengih
maindera rectanglewith
obtained.
equal toAD, can be
b
joinedto
This is sliced and
O F,G ABC, D, as shown. The sliced
be
as Lo gel Fig. 15 to
maining square issuitably so
a
EAbe sarrOw strip and together
with lengthput A D,: which is
then
sOUAIE Wilelnglbeh hegakes han imported
BAlo B.SL. I. 53bizes thrice
square, the breadth, the ãgantu or
Zulbasütra Geometry 37

againjoined to A F, F D,. This procedure is not merely empiri

cal but also highly unsatisfactory, since there is no guarantee
that the remainder square always yields a rectangle of the
requisite length.
ipastamba's commentator Sundararja, explaining this sätra,
gives a purely geometrical and exact,construction.

(Producing the sides of the square castward to the desired length

of the lateral side, one should draw the north-eastern diagonal.
The part of the transverse side to the
F H E north of the point where the diagonal
cuts it is to be discarded and its southern
part is to be made the transverse side of
D C the rectangle. That will be the rectangle.)
Let ABCD be the given square.
Produce AD and BC to F and E so
the
that A F- BE= the required side of
rectangle
rectangle. Complete the
diagonal BF
A BEF and join the
G draw a st.
A B cutting CD in G. Through
Fig. 16 line IH parallel to the sides of the square.
Then IBEH is the required rectangle.
For, FA B=FEB
IG B=GCB
FDG=FHG
and
D=rect. G CEH
Hence rect. A IG BCG+rect. GCEH
IBEH=rect. I
.". rect, BCG+ rect. AIGD
-rect. I
=Sq. ABCD
where the given side of the rectangle is greater
Here the case
only is dealt with. But with slight
than the side of the square case where the side is less
applicable to the
Change the method is
also. easier method.
Baudhäyana and Kätyâyana give an
(K. S. Il. 4)

'Also B. SI. I. 52.

 Geometryin Ancientand Medieval
(Wishingtotransforma square into a rectangle one phoula
38
divide one part again
Indo
in north and east of the other part. and
e middle,
fwo halves tothe
cut
diagonally the
should place together as it
Ifthe figure
place the
one
quadrilateral distribution), Let fits. This is the
isa
the square.. Itis cut BCD h
form two right allong AC to
ABCis again thalved
riangles. T}he
C altitude BE. The twoalong the
BEA and BEC are then halves
moved to the positions DEC
and D G A.
Then, obviously rect. ACFG
=Sq. A BCD. The defect of
this method is that the rectangle
B
cannot be given any desired
Fig. 17 side.

2.10.5. To convert a rectangle or square into a trapezium with the

shorter parallel side given.
Baudhyana deals with this problem.

(Lf one wishes to (B. SI. I. 55)

make a Square or
SIde, one should cut off a rectangle shorter on one
portion by the shorter side.
remainderon should be divided Ihe
attached by the and
If either side),. diagonal, inverted
be cutABCD is
off so the given
that rectangle, let the rectangle A FED
AF=DE=the given shorter side. The re-
E C maining rectangle EFBCIS
to be cut diagonally along
BE and the portion BEC
is to be inverted and attach-
ed to the sidee ALD of the
A
rectangle in the possition
Fig. 18 FB theEAD. Theen DE'EBEis
equivalent trapezium.
Sulbasütra Geometry 39

The Satapatha Brähmana seems to give another method for

this conversion,!

(He measures by 24 angulis. For Gâyatri has 24 letters and agni

is of Gyatrl. Asmuch as is the agni, as much as is its measure,
by that nmuch he measures them. He contracts 4 angulis inwards
on both sides. He stretches 4 angulis outwards on both sides.
as much as he contracts, so much he stretches. Thus he does not
make it exceed (the right measure) Datta
nordoestakes
he make it smaller).
this to mean?
D D C that the face of the square
(A BCD) is shortened on
either side by a small length
(D D'=CC') and the base is
lengthened on either side by
P the same length to A' and B.
Then the area of trapezium
A'B'C'D' =the area of square
ABCD. and obviously so, since
B B APDD' APAA'
A' A
and A QCC AQB B'
Fig. 19
problem.3
other Sulbasütras do not deal with this
The
equivalent rectangle
2.10.6. 7Toconvert a trapezium intoan
converting an
Apastamba tackles the converse problem of It is not given
rectangle.
Is0sceles trapezium into an equivalent a means of finding out
rather as
as a general prescription but Mahavedi.
the area of the trapezium of the

the southern top corner one should drop a perpendicular

(From distance of 12 (padas from
southern botto n corner at a
on the
1$. Br. X 2. 1. 3-4.
Science of the Sulba, p. 91.
(Science of the Sulba, p. 92) that Ap. Sl. XV-9 ff deal with
B.B. Datta says am using the said sütras do not
deal with
s problem, but in the edition I
this problem.
 Ancient and Medieval India
Geor etrj i
40 oul be
e placed inverted at the
Theremoved bit si One should study it thus
the prsthy). rectangle.
is the
northern side. That
joined). is. to convert the trapezium
Tlat ft. and
ABCD (parallel sides 24
D
CB r drawn perpendicular to
30 ft.). away from
AB,B' being 12 padas CB Bis
triangle
PQ, the prthy. The
placed in the position A DD',
then A BCD
Then rect. A B' CD'= trap.
Fig. 20
in aea to a giren
2.10.7. To construct an isosceles triangle equal
square and vice versa.
Conversion ofa square into an equivalent triangle, being necessary
for the construction of the Praugacit, is tackled by all the three
important Zulbasätras and allof them give the same prescription.

(A. SI. XII. 5)!

(Making an area which is double as much as the fire-altar with
the aratnis and prde[as, into a square, one should
draw lines
from the middle point of the eastern side towards the
corners. That is the equivalent prauga (isosceles bottom
D E
triangle).
Let ABCDbea square of twice the
required area. Let E be the middle
point ofCD. E Aand EB are
Then A EB is the required joined.
if the altitude EF is triangie. For
divided into 2 cqualdrawn, the square 1s
rectangles AFED
and FBCE.
B
and
AFE - rect. AFED
Fig. 21 FBE = rect. FBCE
.. Whole
AEB= sq. ABCE.
See also B. SI. I.
56 andK. SI. IV. 5.
Slbasitra Geomery 41

This construction leads to the forula,

area of a triangle base X altitude.
The converse problem is treated by Katylyana c y

(IV 7
Wishingto convert an isosceles triangle into a square onesheould
cutin the middle towards the east. place on the other side in
eted and mampulate by the method of conversion of a
rectangle into asquare.)
G
Let A BCbe the triangle. Let it be cuf into
two halves along the altitude C D Let the
part DBC be applied inverted to tlhe side
A C. Then rect. ADCEABC.
Now the rectangle can be converted into an
equivalent square by the method already
given.
D B
Fig. 22
2.10.8. To construct a rhombus of given area

(Ap.S. XIL 9)
area of the
(Drawinga rectangle of the same area (i.e. of the the middle
lines from
square for the prauga), one should draw middles of the
sides to the
points of the castern and western
of the same
Southern andnorthern sides. That is the rhombus
area).
Let A BCD be a rectangle of twice the
H be
area of the rhombus. Let E, F, G, F G,
D C
E,
the middle points of the sides. Erhombus
GH, HE are joined. Then
H F EFG H- } rect. ABCD.

A B
Fig. 23
"Also B. SI. I. 57 and K, St. IV. 6.
 42 Geometry in AnCient and
2.10.9, To ransfrm a rhomhus into a rectangle
Medieval India
This converse construction occurs in Katyäyana only

Eit is an wbhavatab prauga one should cut

(K.Sl. iv. 8)
the middie and join together as before.) transversely alone
The process is exactly the same as for
the praga. The rhombus is first divided in.
to two isosceles triangles by joining a diagonal
and again into 4 right triangles by cutting
D along their altitudes. The four triangles are
joined together to form a rectangle

Fig. 25
2.11.1. Combination of areas and the con verse
By the application of the theorem of the square of the diagonal
the authors of the Sulbasitras combine any number of squares to
form another square.
For, combining two equal squares i.e. for doubling a square
Apastamba's rule is:
(1. 5)
(The diagonal of the square is the double-maker. ) Hence if a
square is drawn on the diagonal of the given square, it will
produce double the area. The diagonal will therefore be y2 a
where a is the side of the square. It is noteworthy that the
Sulbasttras give a very close approximation to the value
of y 2

(Ap. Sl. I. 5and B. SI. I. 6l-62)4

(The measure should be increased by one-third of itselt,
which again is increased by its one-fourth and diminished by
34 th of that (second) increment. This is the savi[e_a.)
Also K. S. II. 13.