Manasara (English translation)

by Prasanna Kumar Acharya | 1933 | 201,051 words

This page describes “the rules for erecting gnomoms and pegs” which is Chapter 6 of the Manasara (English translation): an encyclopedic work dealing with the science of Indian architecture and sculptures. The Manasara was originaly written in Sanskrit (in roughly 10,000 verses) and dates to the 5th century A.D. or earlier.

Chapter 6 - The rules for erecting gnomoms and pegs

[Full title: The rules for erecting gnomoms and pegs (śaṅku-sthāpana-lakṣaṇa)]

1. After this I shall fully describe the rules for erecting a gnomon[1].

2. At sun-rise the erection of the gnomon should be undertaken.

8-6. In a month of the northern (December 22 to June 21) or the southern (June 22 to December 21) solstice, in the bright or dark fortnight, on the most auspicious day, excepting the full moon (day) and the now moon (day), and at a very auspicious moment, in the morning the gnomon should be erected. Thereafter it should remain there till the evening.

7-9. On the day previous to the erection (of a gnomon), purification of the place (where the gnomon is to be erected) should be carried out: in the middle of the selected site a spot, quadrangular (in shape) and measuring four cubits each, way, (should be made) watery all over (in order to secure the accurate levelling).

10-12. The trees (from the timber) of which the gnomon is stated to be made arc those; kṛtamāla (Cassia fistula), the branch of śamī (a kind of fire-producing) tree, sandal (Sirium myrtifolium), red sandal Caesalpinia sappan), khadira (Acacia catechu), tinduka, (Diosphyros ombryopteris), white milk-tree (Mimusops kanki), or śubhadanta (tooth-tree).

13-14. The length of the gnomon should be one cubit (i.e., eighteen inches) and the width, at the bottom six aṅgulas (of three-fourths inch each); the width at the top-end should be two aṅgulas, (the whole) gradually tapering from the bottom to the top.

15. Its top should be quite circular, smooth and shaped like an umbrella.

16. This is the large type of gnomon, the intermediate one is now described.

17-18. Its length should be eighteen aṅgulas and the width at the bottom fives aṅgulas and the width at the top-end one aṅgula, and the rest should be made as aforesaid.

19-21. The length of the smallest gnomon especially should be twelve aṅgulas, the width at the bottom four aṅgulas and at the top-end one-third aṅgula; as an alternative the length may be nine aṅgulas and its width at the bottom and at the top-end (should correspond), and the rest is stated to be as before.

22-24. In the centre of the selected site the expert geometrician[2] should describe a circle by moving around (a cord of) twice the length of the gnomon (as the radius); and on the centre (of the circle) a gnomon should be fixed.

25-28. In the forenoon (at a certain time) the chief architect) should mark a point; (whore) the shadow from the gnomon (meets) the circumference in the west. In the afternoon (also) a point should be marked as before (i.e., as ia the morning)[3] where the shadow from the gnomon (meets) the circumference in the east. Thereafter the gnomon should be left (to remain) therein.

29-30. The length of the gnomon being divided into ninety-six parts, (and) the apacchāyā [see notes on apacchāyā] being left out of these parts, the (due) east should then be determined.

31-35. In the months of Kanyā (August and September) and Vṛṣa (April and May) there is no apacchāyā. The apacchāyā left out is two aṅgular[4] in the four months, Meṣa (March and April), Mithuna (May and June), Tulā (September and October) and Siṃha (July and August); four aṅgulas in (the months of) Vṛścika (October and November), Āṣāḍha (or Karka, i.e., June and July) and Mīna (February and March); and six aṅgulas in (the months of) Dhanus (November and December) and Kumbha (January and February); and the apacchāyā is stated to be specially eight aṅgulas in (the month of) Makara (December and January).

36-37. The aforesaid aṅgulas should be marked in the shadow to the left and right of the centre[5]; (with) what is left after the deduction of these aṅgulas the due east line should be drawn.

38-39. During the sis months (i.e., northern solstice) beginning with Makara (December 21-22) the shadow declines towards the south and during the sis months (i.e, southern solstice) beginning with Kulīra (June 21-22) the shadow declines towards the north.

40-47. In the shadow facing the east-left the left (point) should be marked; thereafter moving towards the east and right the west-left points should be marked.[6] The architect should leave out the apacchāyā and draw the east-west line. By taking (the cord) through the north direction towards the east region (thus) the extension of the fish should be made and the aṅgula[7] (should be marked) in front. The door (i.e., entrance) of it (fish) should be marked to the south and north of that line; the line drawn joining the head and tail of the fish should be the north-south line [see notes on cardinal points below]. The point should be marked by moving the cord up to the circular orb (i.e., circumference).

48-49. The apacchāyā is (further) specified as it varies during the three parts, of each month, of ten days each.

50-51. In the month of Meṣa (March and April) two aṅgulas (of apacchāyā) should be left ouy during the first ten days, one aṅgula, (part) during the middle ten days and none during the last ten days.

52-53. lathe month of Vṛṣa (April and May) none at all should be left out daring the first ten days, one aṅgula (part) is stated (to be left out) during the middle ten days, and two parts during the last ten days.

54-55. In the month of Mithuna (May and June) two aṅgulas (parts) should be left out during the first ten days, three aṅgulas (parts) during the middle ten days, and four aṅgulas (parts) during the last ben days.

56-58. In the month of Kulīra (June and July) lour aṅgulas (parts) should be left out during the first tea days; during the middle ten days also three aṅgulas (parts) should be left out as that (i.e., the light-shadow); and two aṅgulas (parts) are stated (to be left out) during the last ten days.

59-60. In the month of Siṃha (July and August) two aṅgulas (parts) should be left out during the first tea days, one (part) during the middle ten days, and none during the last ten days.

61-62. In the month of Yuvatī (August and September) none should be left out during the first ten days, one aṅgula (part) should be left out during the middle (ten days), and two aṅgulas (parts) during the last ten days.

63-64. In the month of Tulā (September and October) two aṅgulas (parts) are forbidden (i.e., left out) during the first ten days and three aṅgulas (part) should (also) be left out during the middle (ten days), and four (parts) are known (as forbidden) during the last (ten days).

65-66. In (the month of) Vṛścika (October and November) four aṅgulas (parts) (should be left out) during first (ten days), five (parts) during the middle ten days, and six aṅgulas (parts) during the last ten days.

67-69. In the zodiac (i.e., month) of Dhanus (November and December) six aṅgulas (parts) should be left out during the first ten days and seven aṅgulas (parts) during the middle tea days, and it is (well) known that eight aṅgulas (parts) should be left out during the last ten days.

70-71. In (the month of) Makara (December and January) the wise (architect) should leave out eight aṅgulas (parts) during the first ten-days, seven aṅgulas (parts) should be left out during the middle (ten days), and six aṅgulas (parts) should be left out during the last (ten days).

72-73. In (the month of) Kumbha (January and February) he (the architect) should leave out six aṅgulas (parts) during the first ten days, five aṅgulas (parts) should be left out during the middle (ten days), and four aṅgulas (parts) during the last (tea days).

74-76. In the zodiac (i.e., month) of Mīna (February and March) four aṅgulas (parts) should be left out during the (first) tea days, and three aṅgulas (parts) during the middle ten days, and during the last ten days also two aṅgulas (parts) should be left out,

77-82. The occasions as has been stated (by the ancients) when there is no apacchāyā will now be further specified here: in the aforesaid solar zodiac in Kanyā (August and September) and Vṛṣabha (April and May) daring the other (i.e., last) twenty days should there happen to be a constellation[8] the aforesaid (measures in) aṅgulas (of apacchāyā) should be taken as nil. Knowing this he (the architect) should use the cord (to find out the cardinal points). In these solar months even if those constellations take place (only) occasionally it (apacchāyā) should be left out, (because) the sages have allowed discretion to accept or reject in case of doubt (to the extent of) ten (? two) aṅgulas[9].

88- 84. In accordance with these (rules)[10] the (different) quarters (i.e., the points of the compass) should be determined (and) the cord should be spread therein[11], the due east being (first) determined most perfectly (i.e., accurately)[12].

84-86. Then the north-east is stated (to be found out): the aṅgula (? point) is marked to the north of the point of the east aṅgula (already) marked and the north-east line is drawn from the point (extending) up to the west[13].

87. The due east should be preferred for the building of those who desire salvation (i.e., temples should face due east).

88. The north-east is preferred for (the building of) those who seek enjoyment (i.e., residential buildings should face north-east).

89- 90. That (building) which faces south-east is the source of all evils: therefore, all (kinds of) buildings with face towards the south-east should be avoided.

91. This (instruction) being observed a building is capable of bringing forth an intensive and extensive prosperity.

92. The measurement of the length of the (measuring) cord should conform to the length of the (measuring) rod[14].

93-94. Dividing the threads (of the cord) in three folds the Sūtragrāhin (i.e., measurer or designer) should join them (in the following manner): at first the cord should be of two folds and the third fold (should be joined by taking it round the two-folded cord) by the right side.

95. It (the cord) should be made either of cotton or of jute threads.

96. The (more) accurate dimension in a building can indeed be secured (when it is measured) by the cord (rather than by the rod, in the following manner).

97-100. He (the architect) should move the measuring cord (in order to find out the dimensions of an architectural object) taking it (first) from south-west as middle towards its (connecting) directions (i.e., south, and west), (then) from east to south-east, from east to north-east, from south to south-east, from west to north-west, from north to north-west, and from north to north-east.

101-102. With (this) measuring cord should be measured the architectural objects, such as extensive temples, large residential buildings[15] (for kings and richer people) and (humbler) pavilions (i.e, small buildings[15], both religious and residential), and all kinds of villages, etc, (i.e., inclusive of towns and fortified cities).

108-104. From beyond the extreme end of these points found out by the measuring cord at a distance of a cubit or two around the dimensions (of an architectural object) thus ascertained (wooden) pegs are, as stated (below), fixed[16].

105-108. Two pegs each at the corners (lit. ears) of the central line should be fixed. For the exit of the foundation four pegs should be fixed at the four quarters (i.e., north, east, south and west). Four pegs at the four corners (ears) should be fixed and the corners should be attached (i.e., joined)[17]. This is of great benefit, it should be done.

108-110. The wood with which these pegs are made will be described (now); the trees with (the timber of) which the pegs to be (thus) fixed are made are these: khadira (Acacia catechu), ādimeda (a plant), madhūka (Bassia latifolia), and similarly milk-tree (Mimusops kanki) and others, or the pithy trees.

111-112. The length of these pegs to be fixed should be twenty-one or twenty-five aṅgulas (of three-fourths inch each) and its width should be (equal to) the measure of one’s fist (i.e., about seven inches).

113. Its bottom should be made (pointed) like a needle, but (from above the ground) it should taper gradually from bottom to top.

114-115. The architect and the master standing with face towards the east or north, (each) catching hold of a peg by the left hand and holding a hammer should strike it (the peg) with the right hand, and there should be eight strokes on each (of the pegs).

117-118. At the time of the fixing of the pegs, the Brahmin (priest) should pronounce benediction, and thereafter those who assemble (at the laying function of the foundation) should (also) pray (for the success of the undertaking) with all auspicious sounds.

119-120. After this (ceremonial posting of pegs) the carpenter with the permission of the (chief) architect should in the same way strike all the pegs beginning with south-west corner amidst all auspicious sounds.

Thus in the Mānasāra, the science of architecture, the sixth chapter, entitled: “The rules for erecting gnomoms and pegs.”

Notes [40-47]: Finding out cardinal points by means of a gnomon:

The ordinary mode of finding out cardinal points by means of a gnomon is simple. A gnomon of 24, 18 or 12 aṅgulas in length, 6, 5 or 4 aṅgulas at the base and 2, 1 or 5/4th aṅgula at the top which is shaped like an umbrella, is fixed on the selected ground levelled with water. A circle is described from the bottom of the gnomon with radius twice the length of the gnomon. Two points are marked on the circumference of the circle when the shadow of the gnomon meets it before and after noon. The straight line joining these two points is roughly taken to be the east-west line (vide lines 30, 37, 42, 84). The line which bisects the east-west line would be the north-south line.

The bisecting is done in the usual way. With each end of the east-west line as centre and the length of the line as radius two circles are drawn which intersect each other at two points forming a fish-like common, segment; the straight line joining these points of intersection bisects the east-west line at right angles and indicates the north-south line. The intermediate quarters are found out in the same way by constructing the fish between the points of the determined quarters (see plate VI, fig. I).

The inaccuracy in the precise determination of the east and west points is caused by the variation of the shadow in consequence of declination of the sun during the interval between the two instants in the forenoon and afternoon when the shadow is observed. For the purpose of rectifying the inevitable variation of the shadow apacchāyā is stated to be deducted from the shadow (see note under line 29-30).

The subject has been discussed more or less elaborately by all the other leading authorities of astronomy and architecture, for instance,

  1. Sūrya-siddhānta of Bhāskarācārya (III, 1—51),
  2. Brahma-sphuṭa-siddhānta of Brahmagupta (XIX, 1—20).
  3. Līlāvati of Bhāskarācārya (XI, 1—10; part II, chapter II, section 4).
  4. Siddhānta-śiromani of Bhāskarācārya (VII, 36—29).
  5. Pañca-siddhāntikā of Varāhamihira (II, 10—13, XIV, 1—11, 14—22),
  6. Mayamata (VI, 1—28).
  7. Śilparatna of Śrīkumāra (XI, 1—22),
  8. Kāśyapa-śilpa (I, 60—70).
  9. Vāstu-vidyā (III, 7—10).
  10. Manuṣyālaya-candrikā (II, 1—4).
  11. Vitruvius (Book I, chapter VI, Book IX, chapters IV, VIII, IX)

Footnotes and references:


For dialling, finding out cardinal points, and orientation and planning of buildings.


Literally one who is conversant with the point (bindu).


The morning shadow and the evening shadow meet the circumference of the circle approximately at an equal interval from the noon.


Aṅgula in these lines and Mātra in lines 56, 66 and elsewhere are indiscriminately used for Aṃśa (part) or degrees. For different senses in which aṅgula is used see the writers ‘Indian Architecture’ (pages 35, 77, 121, 122) and his Dictionary of Hindu Architecture’ under aṅgula.


Bindu, which means a ‘point’, it may also imply the centre of a circle or a point of intersection of any two lines.


Apparently what is intended to be implied (in lines 40-41) is this: in the shadow moving to the east by the left the left point should be marked, thereafter moving towards the west, opposite the Tight, i.e., left, the right point should be marked.


Obviously it seems to imply a ‘point’, tut it may indicate the measure of apacchāyā.


Which is assigned to the sixth (Kanyā) and second (Vṛṣa) zodiacs.


The rendering of last two lines (81-82) is tentative as it contains a grave objection, namely, when the maximum correction can be only eight it would be useless to allow to exercise discretion to the extent of ten, although instances of such an incongruity are not rare in the Mānasāra and other texts, It should be noticed that the lines 81—83 though preserved by all the other Mss. have been altogether left out by the Codex architypus [archetypus?]: they need not be taken into consideration at all.


Which include the aforesaid modification, exception and exercise of discretion.


That is, by means of a cord with which, the necessary circles and the required lines are drawn.


Because, otherwise, the determination of the other quarters would not be precise, as their accuracy depends upon the perfection of the east-west line. For the full details see note underlines 40- 47.


For details see note under 40—47.


Eight rods make one cord, see chapter II, 53.


Trigṛha seems to imply residential houses with three courts and courtyards, while Vimāna would mean huge temples with, five courts and Maṇḍapa would refer to humbler buildings, both religions and residential, with a single court-yard (see the writer’s Indian Architecture, pages 51-52, 47, 48, 53-54, and also his Dictionary of Hindu Architecture under these terms. In the Mṛcchakatika, (act IV) a palace possessing eight court-yards is described in detail, but tridhātu-śaraṇa is mentioned as a three-storeyed building in the Ṛgveda (see Indian Architecture, pages 32, 6).


In order to fasten, strings for laying the foundation.


See plate VI, fig. 4.

Like what you read? Consider supporting this website: