The orientation of buildings : or, Planning for sunlight

Preface

The purpose of this book is to set forth the principles which ought to govern the planning of buildings with respect to sunlight, a subject to which very little attention has been given.

Several years ago, in an essay on hospital construction,1 I wrote as follows: "To study properly the question of sunlight, a sun plan of the buildings must be drawn, and their positions considered with respect to the shadows they cast upon each other and upon the ground." This statement describes very well the general method of study which I have followed in my investigations, the results of which are now for the first time presented to the public in a complete form.

I had begun my study of orientation with hospitals especially in mind, but I soon realized that the general principle of planning with reference to sunlight was of fundamental importance in the design of all buildings, and especially in the planning and laying out of cities. In this connection I may mention that a series of diagrams made by me, at the request of a committee interested in securing new legislation regulating the height of buildings in Boston in 1904, was of great service in showing the effect of tall buildings

1 "Small Hospitals," by A. Worcester, M.D., and "Suggestions for Hospital Architecture," by William Atkinson, Architect, New York, John Wiley & Sons, 1894.

v

VI PREFACE

in overshadowing and shutting out the sunlight from the streets. Another series of street diagrams, first shown at a lecture given by me before the Society of Arts, at the Massachusetts Institute of Technology, has been reprinted by permission in a recent English book on city planning. Some of my earlier studies in orientation, originally published in the National Hospital Record, have been twice reprinted in The Brickbuilder magazine, and also embodied, by permission, in a recent American book on hospital construction.

All of which has encouraged me to believe that a more complete presentation of the subject, in book form, would not be without interest to the public.

In my first chapter I have included so much of the elements of astronomy as is necessary to a clear understanding of the apparent motion of the sun, and the variations in the angles of sunlight at the different seasons. I have also described the method of the stereographic projection, by which the angles of sunlight may easily be obtained, for any season of the year, and for any latitude.

The second chapter deals with the distribution of sunlight upon the exterior of buildings, and its admission to the interior, through windows.

In this chapter I have developed the method of the "shadow curve" and the "area of complete shadow," an application of the principles of descriptive geometry to the recording of transitory occurrences, which, as far as I am aware, is new. General principles for the planning and placing of buildings are given as far as it has seemed desirable. It is, however, to be understood that, for the

PREFACE Vll

best results, each case must be studied as a separate problem, with reference to local conditions, and especially with reference to the latitude of the place. In connection with the study of windows an account is given of my experiments with the "sun box," an apparatus devised by me to test the practical effect of different window exposures.

My third chapter is devoted to hospitals. In it I have discussed the vexed question of the best orientation for hospital ward pavilions and have ventured to make recommendations in this regard at variance with common practice. I have also presented a plan for a new type of hospital building especially designed to meet the needs of modern medical treatment.

The last chapter is concerned with the distribution of sunlight in streets, as affected by their direction and width, and the height of the buildings upon them. In an appendix I have given in full the building law of Paris regulating the height of buildings and a synopsis of the regulations of some American cities in this matter.

While the working out of the diagrams and the calculation of the tables has been a matter of pure mathematics, admitting but one result, the conclusions to be drawn from them are to some extent a matter for individual judgment.

It is therefore fitting that I should give here a statement of the premises on which my recommendations are based.

I have assumed that it is desirable, in our climate, that all buildings in which human beings dwell or work should have all of their exterior walls exposed to direct sunlight at some time during the day throughout the year, and that the surfaces of streets, alleys, areas, courtyards, and other

Vlll PREFACE

spaces in and around buildings should also be exposed as much as possible to the action of direct sunlight.

In regard to windows, I have assumed that as much direct sunlight as possible is desirable through them during that period of the year in which they are customarily kept closed. On the other hand, during the hot season, when windows may be open day and night, and the sun-purified air brought in from outside by natural means, I have assumed that direct sunlight through windows is rather to be avoided than sought after.

The function of sunlight in promoting healthy conditions, and its use as a therapeutic agent, may only be authoritatively stated by sanitarians and medical men who have given special study to this question, and I can do no more than refer those who may wish to pursue this branch of the subject to the work of the German and Danish investigators, a full account of which may be found in Luft-und Sonnenbdder , by Dr. Julian Marcuse, Stuttgart, 1907.

The diagrams for this book have been engraved by the wax process from drawings made by me, and the few which have appeared in former published articles have been carefully revised and redrawn.

Boxford, Mass., October, 191 1.

Chapter I

Sunlight a requisite for healthy buildings. — An elementary knowledge of astronomy necessary for intelligent sun-planning. — The angles of sunlight at the different seasons and in different latitudes. — Calculation of the angles of sunlight by the stereographic projection. — By spherical trigonometry. — Table of sunlight angles.

Chapter II

Shadows of the cube. — Orientation of the Swiss house. — Shadow curves of the cube. — Theory of the area of complete shadow. — The area of complete shadow as applied to the study of fundamental types of building plan. — Sunlight admitted by windows. — The visual angle of windows. — Quantity of sunlight admitted by windows. — Heating effect of sunlight. — The solar constant. — The sun box. — Sun-box records.

Chapter III

The orientation of ward pavilions. — Different views upon the subject. — Recommendations of the author. — The typical ward pavilion. — Its unsuitability to modern conditions. — A new type of ward needed. — Description of the pyramidal type of ward.

Chapter IV

Angles of sunlight in streets. — Sunlight curves in streets. — The orientation of streets. — Horace Bushnell's theory. — ■ The height of buildings. — European building regulations. — The law of ancient lights. — Building law recommended by the author. — The skyscraper and the street.

APPENDIX A

Sun tables for London and New Orleans.

Building laws of Paris regulating the height of buildings.

APPENDIX C

Building laws of American cities regulating the height of buildings.

LIST OF ILLUSTRATIONS

Chapter I

Fig. Page

i. Cross sections of the visible celestial sphere 3

2. Apparent orbits of the sun at the different seasons 6

3. Paths of the sun at intermediate periods 7

4. Stereographic projection of the visible celestial sphere 9

5. Construction of the stereographic projection. — First step 12

6. Construction of the stereographic projection. — Second step 14

7. Construction of the stereographic projection. — Third step 16

Chapter II

8. Shadows of the cube 20

9. Shadows of the cube 21

10. Shadows of the cube 22

11. Plan of typical Swiss dwelling 23

12. Shadow curves of the cube. Winter solstice 25

13. Shadow curves of the cube. Winter solstice 26

14. Shadow curves of the cube. Equinoxes 27

15. Shadow curves of the cube. Equinoxes 28

16. Shadow curves of the cube. Summer solstice 29

17. Shadow curves of the cube. Summer solstice 30

18. Areas of complete shadow: single straight block 32

19. Areas of complete shadow: L plan 33

20. Areas of complete shadow: U plan 34

21. Method of obtaining the area of complete shadow 36

22. Shadow curves of cube and prism at winter solstice 37

23. Maximum areas of complete shadow for the L and U plans 38

24. Effect of increase of height upon the area of complete shadow 39

25. Good and bad arrangement of L 41

26. Application of the test of the area of complete shadow 42

27. Shadows of the single straight block. Axis N. and S 44

xi

Xll LIST OF ILLUSTRATIONS

Fig. Page

28. Shadows of the single straight block. Axis E. and W 45

29. Shadows of the single straight block. Axis N. E. and S. W 46

30. Visual angle of ordinary window 47

31. Visual angle of mediaeval window 48

32. Wall section with beveled piers 49

33. Window illumination: winter solstice 50

34. Window illumination: equinoxes S1

35. Window illumination: summer solstice 52

36. Cross sections of sunlight prism 53

37. Change in area of sunlight prism: winter solstice 55

38. Change in area of sunlight prism: equinoxes 56

39. Change in area of sunlight prism: summer solstice 57

40. Windows : obstructed outlook 59

41. Obstructed outlook: stereographic projection 61

42. Change in area of sunlight prism: obstructed outlook 62

43. Cross section of sun box 64

44. Photograph of sun box 65

45. Sun box records 76

46. Sun house 77

Chapter III

47. Ward pavilion: open-ended type 80

48. French method of hanging outside blinds 84

49. Economy of two-story type of pavilion 86

50. Section of ward with ridge ventilation 88

51. Types of ward pavilions 89

52. Grouping of ward units: Virchow Hospital 91

53. Grouping of ward units 92

54. Grouping of ward units 93

55. Grouping of ward units 95

56. Elevation of Virchow ward unit 97

57. Pyramidal type of ward construction 98

58. Pyramidal ward unit: first-floor plan 99

59. Pyramidal ward unit: second-floor plan 100

60. Pyramidal ward unit: third-floor plan 101

61. Pyramidal ward unit: shadow diagram 102

LIST OF ILLUSTRATIONS xm

Fig. Page

62. Pyramidal ward unit: shadow diagram 103

63. Pyramidal ward unit: shadow diagram 104

64. Pyramidal ward unit: general plan 1 06

65. Pyramidal ward unit: elevation 107

Chapter IV

66. Angles of sunlight in streets in

67. Angles of sunlight in streets 112

68. Angles of sunlight in streets 113

69. Method of obtaining sunlight curves 114

70. Sunlight curves: streets 116

71. Building law of Paris 119

72. Typical cornice section 122

73. Proposed building law 123

74. The skyscraper and the street 124

THE ORIENTATION OF BUILDINGS

Chapter I

Unquestionably one of the first requisites for a healthy building is abundance of sunlight. Not only the exterior wall surfaces of buildings, but also the surfaces of the ground around them, should have the direct rays of the sun for as long a time as possible each day.

"Second only to air, is light and sunshine essential for growth and health; and it is one of Nature's most powerful assistants in enabling the body to throw off those conditions which we call disease. Not only daylight, but sunlight; indeed, fresh air must be sun- warmed, sunpenetrated air. The sunshine of a December day has been recently shown to kill the spores of the anthrax bacillus." (Healthy Hospitals, Sir Douglas Gal ton, Oxford,

To secure sunlight in fullest measure requires careful and intelligent planning with this end in view. It is necessary for such a study to have at hand a table, giving the angles of sunlight at the different hours of the day and at the different seasons of the year, for the particular latitude in question.

In this chapter I shall describe one method by which such a table may be prepared.

In all of the operations of practical astronomy, as in the calculation of position of ships at sea, or in determinations of latitude and longitude upon land, it has been found best to go back to the conceptions of the first astronomers, who imagined the earth to be the center of the universe, and the celestial bodies to revolve around it.

And thus has survived the ancient fiction of the " celestial sphere."

Viewing the heavens on a starry night, the whole firmament seems slowly to revolve, successively bringing into view, above the eastern horizon, one constellation after another.

If, by magic power, the sun's light could be dimmed so that the stars should be visible in the daytime, he would appear, like them, to be fixed in the celestial sphere, and to turn with the constellations in their uniform diurnal motion around the pole.

But if we could extend our observations over a period of several weeks, we should observe that the sun was slowly changing his position among the stars, passing in the course of a single year through the successive constellations of the zodiac, in summer north of the celestial equator, in winter south.

But this change in the apparent position of the sun is so slow that for any single day it may be disregarded, and his position for that day considered as fixed in the celestial sphere.

NEW ORLEANS

Fig. i. — Cross sections of the visible celestial sphere, showing the path of the sun at the solstices, and at the equinoctial periods, for different latitudes.

Fig. I shows in cross section that part of the celestial sphere which is above the horizon, at the latitudes respectively of London (Lat. 5i°-3o' N.), Boston (Lat. /\2°-22' N.), and New Orleans (Lat. 3o°-o' N.).

In tnese diagrams H'H represents the plane of the horizon; 0 the position of the observer; P the celestial north pole; Z the zenith,- and SS\ EO, and WW the apparent paths of the sun at the periods of the summer solstice, the equinoxes, and the winter solstice, respectively.

It is evident that the altitude of the sun, at noon, at the periods of the year referred to, may be obtained directly from the diagrams, being given by the angles HOS, HOE, and HOW, for the summer solstice, the equinoxes, and the winter solstice,. respectively.

It will be observed that these angles are less in the more northerly latitudes, and that the path of the sun inclines more and inore toward the horizon.

This decrease in the altitude of the sun is accompanied with an increasing divergence between the extreme points of sunrise and sunset, so that the days in summer are much longer, and in winter much shorter, in the countries of the far north, than in those which are near the equator.

At latitude \2°-o' N. (approximately the latitude of Boston, Mass.) the sun rises on the longest day of the year about 32I0 north of east and sets at an equal angle north of west, reaching at noon an altitude of yi°-2yr above the horizon.

On the shortest day of the year he rises about 32§° south of east and sets at an equal angle south of west, reaching at noon an altitude of only 24°-33' above the horizon.

At the two periods of the year when day and night are of equal length he rises in the east and sets in the west, reaching at noon an altitude of 48°-o/ above 'the horizon.

The perspective diagrams of Fig. 2 will give the student an easily-remembered mental image of the path of the sun at these periods.

The horizontal circle represents the horizon; the inclined circle the path of the sun, and the diverging lines the direction of the sun's rays at the different hours of the day.

It will be noted that at the period of the equinoxes the trace of the sun's rays describes a plane; at the period of the summer solstice a hollow cone, and at the period of the winter solstice a convex cone.

His path at intermediate periods may be pictured by the aid of the following diagram (Fig. 3) which gives his position at intervals of one month apart throughout the year.

It will be observed from this diagram that the path of the sun during the four months from April 21 to August 21 resembles more nearly his path at the summer solstice than his path at the equinoxes, and similarly his path during the four months from October 21 to February 21 more nearly his path at the winter solstice than his path at the equinoxes.

This must be borne in mind in studying the various shadow diagrams which are given later in this book. Those which are drawn for the period of the winter solstice may be taken as typical of the four months from October 21 to February 21; those which are drawn from the period of the summer solstice of the four months from April 21 to

Sunrise ii m t

"Sunset

Sunrise E

XI HI

W Sunset

11 Hi

Sunset

Fig. 2. - Perspective diagrams showing the apparent path of he sun, «dtib andes of sunlight at the different hours of the day, for Lat. 42 -o N. The upper dLtim i^ra for the summer solstice, the middle diagram for the vernal and autumnal equinox, and the lower diagram for the waiter solstice.

August 21 ; and those which are drawn for the period of the equinoxes of the two months from February 21 to April 21, and the two months from August 21 to October 21.

The position of the sun with respect to the observer is generally expressed in terms of azimuth and altitude.

Fig. 3. — Cross-section of the visible celestial sphere showing the path of the sun at periods one month apart throughout the year. Lat. 42°-o' X. Actually the declination of the sun on May 21 does not exactly coincide with his declination on July 21, although it is so represented in the diagram. A similar observation applies to the other dates which are grouped in pairs. The differences, however, are so slight that it would be difficult to represent them at the scale at which the drawing is made.

The latter term requires no explanation, but the meaning of "azimuth" may not be so generally understood. It may preferably be explained by an example rather than a definition.

Imagine a stick set upright in level ground in the sunlight. The deviation of the shadow cast by the stick from a true north and south direction is the azimuth of the sun at that moment.

Knowing the distance of the sun north or south from the equator (which information may be obtained from the almanac) the azimuth and altitude for any particular day may be calculated by spherical trigonometry.

The desired data may also be obtained very simply and easily, and with sufficient accuracy for our purposes, by the stereographic projection.

To one who understands perspective, the stereographic projection presents little difficulty, as it is virtually the method of linear perspective applied to the representation of the sphere, but with this difference, that the drawing when completed is viewed from behind the picture plane, instead of in front of it, as in ordinary perspective.

It possesses two properties which make it especially useful; the first being that all circles of the sphere are projected as circles or as straight lines, and hence may be drawn with compasses and ruler; 'and the second being that the angle made by the crossing of two circles upon the surface of the sphere is the same as the angle made by their projections.

The statement that by the aid of the stereographic projection one may, with a few hours' labor, construct a diagram which will give the position of the sun at each hour of the day for any period of the year desired, and for any latitude, should be sufficient to induce the student to master its principles.

Such a diagram, drawn for latitude 42°-o' N., is shown in Fig. 4.

TEE ASTRONOMICAL DATA

This is a stereographic projection of the celestial sphere,

taken upon the plane of the horizon, which is represented

by the circle N, E, S, W, these letters being placed at the

four cardinal points.

N w

Bl

-n

FlG. 4. — Stereographic projection of the visible celestial sphere, upon the plane

of the horizon.

The circular arc WE is the projection of the celestial equator, which is the path of the sun at the period of the equinoxes.

The arc of considerably greater curvature to the north is the projection of the tropic of Cancer, which is the path

IO ORIENTATION OF BUILDINGS

of the sun at the summer solstice, while the more flattened and shorter arc to the south is the projection of the tropic of Capricorn, which is the path of the sun at the winter solstice.

The twelve circles converging toward the upper part of the diagram are the projections of the celestial meridians, or hour circles, 150 apart.

These hour circles may be conceived of as a gigantic cage or framework, fixed in position, and serving as a system of celestial verniers, to mark the passage of the heavenly bodies, which are carried past them with the revolution of the sphere.

The passage of the sun across the successive hour circles marks the hours of the day as shown by a sundial. Consequently the intersection of any hour circle with the circle representing the path of the sun is the stereographic projection of the sun's position for the corresponding hour and period of the year.

And from this projection the azimuth and altitude may readily be found.

For instance, the dotted line OB drawn through the intersection of the 11-0' clock hour circle and the celestial equator gives the azimuth of the sun at 11 a.m. {solar time) at the period of the equinoxes, and the dotted line OB 1 drawn through the intersection of the 1-0'clock hour circle and the tropic of Capricorn gives the azimuth of the sun at 1 p.m. on December 21st, and the dotted line OB2 drawn through the intersection of the 6-o'clock hour circle and the tropic of Cancer gives the azimuth of the sun at 6 p.m. on June 21st.

THE ASTRONOMICAL DATA II

The altitude is obtained by a simple construction. For example, the angle EOD is the altitude of the sun at II A.M. at the period of the equinoxes and is found by measuring off on the line OE the distance OC equal to OA and drawing the line SC intersecting the enclosing circle at the point D.

To explain more fully the construction of the diagram, an example will be given and worked out.

Let it be required to find the position of the sun at 10 a.m., solar time, April 16, Lat. 30°-o' N.

Draw a cross section of the visible celestial sphere, as shown in the upper part of Fig. 5.

HH is the horizon, 0 the position of the observer, and Z the zenith. Through O draw the line POP' making an angle of 30°-o' with the horizon.

P is the celestial north pole, for it is shown in astronomy that the altitude of the pole is equal to the latitude of the place.

Draw OE at right angles to POP.

It represents the celestial equator.

From the point E lay off the arc ES equal to the declination of the sun on the date required. This we find from the almanac (for 19 10) to be 9°-59' for April 16.

Draw SM parallel to OE. It represents the path of the sun above the horizon at this period.

The projection is made upon the plane of the horizon, and the station point is upon the surface of the sphere vertically below the zenith, at N.

Any point upon the surface of the sphere is projected

Fig. 5. — Construction of the stereographic projection, first step. The upper part of the diagram is a cross section of the celestial sphere; the lower part its projection upon the plane of the horizon.

by joining it to the station point by a straight line and the point in which this line pierces the picture plane HH is the projection required.

The enclosing circle is the horizon, which is drawn without change since it lies in the plane of the projection.

The circle MS may now be projected.

It is evident that M' and M' are the points at which this circle cuts the horizon, and that S' is the projection of the point S.

Through these three points draw the arc of a circle.

It is the projection required, for it is a theorem of the stereographic projection that all circles of the sphere are projected as circles or portions of circles, with the exception of those which pass through the station point, which are obviously projected as straight lines.

An example of the latter is the meridian or 12-o'clock hour circle, which is projected as the straight line H'S'H'.

To find the point upon the arc M'S'M' where the sun is at 10 a.m. it is necessary to project the io-o'clock hour circle.

(For the sake of clearness the operation is shown in a separate diagram, Fig. 6.)

Since all the hour circles pass through the north and south poles, we have at once, in the projections of the poles (at PP and P'P'), two points of our required projection.

It remains to find the center.

The line LT, equidistant from PP and P'P', contains the centers of all circles passing through those points.

1 1

Is

/ s

S S

FlG. 6. — Construction of the stereographic projection, second step.

To find the point upon this line which is the center of the projected 10-0'clock hour circle, we avail ourselves of the second theorem of the stereographic projection, which is that the angle made by two circles upon the surface of the sphere is the same as the angle made by their projections.

Now the 10-0'clock hour circle makes an angle of 300 with the noon circle, or meridian, where it crosses the latter at the poles, one hour being equal to 150.

Consequently a line drawn through PP, the projection of the north pole, and making an angle of 300 with PP P'P', the projection of the meridian, is a tangent to the projection of the 10-0'clock hour circle, and establishes the center of the latter at once, at the point T, upon the line LT.

The circle may now be drawn.

Superposing the two circles thus obtained in one diagram (Fig. 7), their intersection at A is the projection of the sun's position and the angle H'OX is the true bearing or azimuth of the sun required.

The altitude may be found by a secondary construction. (Lower diagram of Fig. 7.)

It is evident that the straight line XAO is the projection of a circle, vertical to the plane of the horizon, and passing through the zenith and the sun.

The sun's altitude is measured upon this circle, upward from the horizon.

Let us imagine this circle to be revolved into the plane of the horizon, thus bringing the station point to the position N and the zenith to the position Z.

Fig. 7. — Construction of the stereographic projection, third step. (The point P referred to in the test is the projection of the north pole, to the left of 0, on the dotted line E'H'.j

The sun lies at some point upon this circle. To find this point draw the line NA cutting the circle at K. It is the point required, and XOK is the altitude of the sun required.

By following the above method and making the projection at a large scale the angles may be found with sufficient accuracy for the purposes of the architect.

To obtain the result by calculation involves the solution of the spherical triangle POA of which the two sides PA (the north polar distance of the sun), PO (the co-latitude of the place), and APO (the assumed hour angle) are known. (Upper diagram of Fig. 7.)

Solving for the angle A OP and the third side AO gives us H'OX, and XA the azimuth and altitude respectively.

In the following table is given the azimuth and altitude of the sun for 420 north latitude at each hour of the day for the typical periods of the year.

TABLE I

Hour angle.

6oc

Winter solstice.

Equinoxes.

Azimuth. , Altitude. Azimuth. Altitude.

4-17

Summer solstice.

Azimuth. Altitude.

5°-9'

37°-23'

Sunrise and sunset.

EXPLANATION OF TABLE.

The first column gives the solar time expressed in hour angles, one hour being equal to 15 degrees, reckoning either way from noon. For instance, the hour angle 30 degrees corresponds to 10 a.m. or 2 p.m., the hour angle 45 degrees to 9 a.m. or 3 p.m., and so on. Azimuth is east of south for the forenoon and west of south for the afternoon.

No corrections have been made for refraction, the effect of which is to slightly increase the altitude when the sun is near the horizon.

This table is the basis on which all of the diagrams in this book have been constructed.

Similar tables, computed for the latitudes of London and New Orleans, will be found in the appendix.

Chapter II

As the first example of the application of the data obtained in the preceding chapter, let us consider the shadows cast by two cubes, one of them placed with its four sides facing the cardinal points, and the other with its diagonal upon the meridian.