Calendars and Astronomic Ceilings

THE ORIENTAL INSTITUTE OF THE UNIVERSITY OF CHICAGO
STUDIES IN ANCIENT.ORIENTAL CIVILIZATION" NO. 26

THOMAS GEORGE ALLEN, EDITOR" ELIZABETH B. HAUSER, ASSOCIATE EDITOR
oi.uchicago.edu
oi.uchicago.edu
THE
CALENDARS
OF ANCIENT EGYPT
BY RICHARD A. PARKER
THE ORIENTAL INSTITUTE OF THE UNIVERSITY OF CHICAGO
STUDIES IN ANCIENT ORIENTAL CIVILIZATION NO. 26
THE UNIVERSITY OF CHICAGO PRESS CHICAGO ILLINOIS
oi.uchicago.edu
Internet publication of this work was made possible with the
generous support of Misty and Lewis Gruber
THE UNIVERSITY OF CHICAGO PRESS, CHICAGO 37
Cambridge University Press, London, N.W. 1, England
W. J. Gage & Co., Limited, Toronto 2B, Canada
Copyright 1950 by The University of Chicago. All rights reserved
Published 1950. Composed and printed by THE UNIVERSITY OF
CHICAGO PRESS, Chicago, Illinois, U.S.A.
oi.uchicago.edu
PREFACE
Early in the course of my work on the text volume to Medinet Habu III: The Calendar, a project upon which Dr.
Harold H. Nelson and I intend to collaborate, it became evident to me that I could not successfully grapple with
the problems of Ramses III's temple calendar without a thorough investigation into all the calendarial phenomena
of ancient Egypt. Once started, my own predilection for the subject led me farther and farther, so that what was
originally intended as a page or two of footnotes has grown to the proportions of the present volume.
Not all of what I shall present in these pages is new. My obligations to those chronological giants, Brugsch,
Meyer, and Borchardt, are manifold; and I have ransacked the literature in order to avoid claiming as my own
what had long ago been proposed. For some earlier propospls I have more proof to offer and may convince where
others failed. Other propositions are mine and must stand on their own merits.
Calendars and chronology are not of themselves difficult subjects, but they are frequently made so by the as-
sumption of their devotees that everyone understands always what they are talking about. I intend to take the op-
posite extreme and shall assume almost no special knowledge on the part of my readers. I shall not shrink from
extended explanation or from repetition whenever understanding may be furthered. In the Introduction I shall pre-
sent some of the basic astronomical and calendarial concepts with which we shall have to deal in later pages.
Practiced chronologers may go on to the first chapter without concern, but I would urge all others to read the
Introduction carefully.
In the following pages I hope to demonstrate that the Egyptians had three calendars, two lunar and religious,
one civil. I shall begin with a consideration of the lunar day and month, pass on to an analysis of the later lunar
calendar, then discuss the probable nature of the early lunar calendar, and, finally, suggest a possible origin for
the civil calendar. In three excursuses I shall offer solutions to various problems which arose naturally out of
the calendarial material.
The end product of our investigations will be, I hope, to persuade the reader to accept my present conviction
that the lunar calendar is an essential key to a proper understanding of Egyptian festivals and thus in some meas-
ure of Egyptian religion.
In the preparation of these studies I have had the inestimable advantage of frequent discussion with my colleagues
in the Oriental Institute. Professor Henri Frankfort permitted me to present my views for the criticism of his
seminar. Professors John A. Wilson and Keith C. Seele and Dr. G. R. Hughes read the manuscript in preliminary
form and made many valuable suggestions. Professors Otto Neugebauer of Brown University and William F. Al-
bright of Johns Hopkins University gave penetrating criticism. My debts to Dr. Abd el-Mohsen Bakir of the Cairo
Museum and to Professor G. Posener are acknowledged on the appropriate pages. Dr. T. George Allen has given
me welcome editorial criticism. The drawings are the expert and painstaking work of Sue Richert. To all of them
and to others not named I am deeply grateful.
RICHARD A. PARKER
Brown University
May, 1950
oi.uchicago.edu
oi.uchicago.edu
TABLE OF CONTENTS
LIST OF ILLUSTRATIONS.................................... ix
LIST OF TABLES.......... ....... xi
LIST OF ABBREVIATIONS.................................. xiii
INTRODUCTION ......................................... 1
2-3. The Celestial Sphere. 4. The Ecliptic. 5. The Vernal Point. 6. The Moon.
SS7-8. Lunar Months. 9. The Synodic Month. 10. The Phases of the Moon. ll-16.
New Crescent Visibility. 17. Old Crescent Visibility. 18. Sequence of Lunar Months.
19. Full Moon. 20. Tables for Calculating the Moon. 21. Heliacal Rising of Sirius.
22. The Civil Calendar. 23-24. Other Calendars
I. THE BEGINNING OF THE EGYPTIAN LUNAR MONTH.......9
25. The Problem. 26-31. Earlier Views. 32-35. The Beginning of the Day.
36-43. The Names of the Days of the Month. 44-48. The Astronomical Basis for the
Names of the Days of the Month. 49-64. The Lunar Calendar of Papyrus Carlsberg 9.
65-107. Double Dates Involving Lunar Days. 108. Conclusion
II. THE LATER LUNAR CALENDAR............................... 24
109. Foreword. 110-22. The Completion of the Cyclic Calendar. 123-28. The Rule
Governing Intercalation and the Position of the Intercalary Month. 129-30. The Lag be-
tween Lunar and Civil Months. 131-39. The First Schematic Calendar. 140-41. The
Later Lunar Calendar
III. THE ORIGINAL LUNAR CALENDAR............................. 30
142-49. Earlier Theories. 150-51. The Proposed Original Calendar. 152-54. The
Primitivity of a Lunistellar Calendar. 155-57. Sothis and the Inundation. 158-63. The
Moon and Sothis in Texts. 164-75. The Meaning of wp rnpt and Its Equivalence to prt-Spdt.
176-81. The Lists of Feasts in the Mastabas of the Old Kingdom. 182-85. Middle
Kingdom Dates of the w3-Feast. 186-87. The Temple Year at Illahun. 188-219. The
Ebers Calendar and the Fixed or Sothic Year. 220-2 3. The Astronomical Ceiling in the
Tomb of Senmut. 224-25. The Astronomical Ceiling of the Ramesseum. 226-37. The
Names of the Months. 238-41. The Feast of Mmn in the Ptolemaic Period. 242-51. The
Feast of Renutet in the Esna Calendar. 252-53. Conclusion
IV. THE CIVIL CALENDAR .. .................................... 51
254-55. The Problem before 1938. 256-63. The Problem after 1938. 264. The
Problem at Present. 2 65-72. The Proposed Solution. 273-80. The Fifty-nine
Divinities of the Dual Year. 281. Conclusion
vii
oi.uchicago.edu
oi.uchicago.edu
LIST OF ILLUSTRATIONS
Figures
1. The Celestial Sphere with the Earth at Its Center ........ ...................
2. The Celestial Sphere with the Sun at Its Center and the Earth in Its Orbit around the Sun,
to Demonstrate the Ecliptic........ . . ................ ..........
3. The Vernal Point, Where the Ecliptic, Going from South to North, Crosses the Equator . .
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
The Sidereal and the Synodic Months . . . . . . . . . ..
The Variation in the Length of the Synodic Month . . . . . . . . . .
The Effect of the Anomaly of the Moon on Crescent Visibility . .
The Effect of the Obliquity of the Ecliptic on Crescent Visibility .
The Effect of the Latitude of the Moon on Crescent Visibility . . . . .
The Variability of Time between Conjunction and Full Moon . . . . .
The Beginning of the Lunar Month . . . . . . . . . . . . . . . .
The Possible Range of Full Moon . . . . . . . . . . . . . . . . .
Possible Relation between Lunar and Civil Months . . . . . . . . .
The Mean Relation between Lunar and Civil Months . . . . . . . . .
The Proposed Regulation of the Lunar Calendar . . . . . . . . . .
Ivory Tablet of the 1st Dynasty . . . . . . . . . . . . . . . . . .
The Ebers Calendar.. . . . . . . . . . . . . . . . . . . . . . .
The Calibrations on the Interior of the Water Clock of Amenhotep III .
18. The Fragment of the Geographical Papyrus of Tanis Which Names the Last Month of
the Year wp rnpt . . . . . . . . . . . . . . . .. ........ . . . ..
19. The Original Lunar Calendar as Depicted in the Ramesseum Astronomical Ceiling . . . . .
20. Concurrence of Lunar and Civil Years at Installation of Latter and Shift of Civil Year
after Fifty Years . . . . ..... . . . . . . . . .. . . .. . . . . . . . .
21. The Last Years of the 12th Dynasty .. . . . . . . . . . . . . . . . . . . . . ..
Plates
I. Astronomical Ceiling in the 18th Dynasty Tomb of Senmut at Deir el-Bahri
Courtesy Metropolitan Museum of Art, New York
II-III. Astronomical Ceiling in the Ramesseum
Courtesy Oriental Institute, University of Chicago
IV-V. Astronomical Frieze in the Ptolemaic Temple of Edfu
Brugsch, Monumens de l'Egypte (Berlin, 1857), Pls. VII-X 1
VI A. A Late Glazed menat in the Cairo Museum
Courtesy Cairo Museum
VI B. Papyrus Fragment Cairo 58065 from Illahun
Courtesy Cairo Museum
ix
.. . 3
. . . . 14
41
44
54
69
oi.uchicago.edu
oi.uchicago.edu
LIST OF TABLES
1. Conversion of Julian into Gregorian Dates....8
2. The Days of the Lunar Month............................... 11
3. The 25-Year Cycle as Given in Pap. Carlsberg 9 .. ......................... 15
4. The Double Dates Entered in the 25-Year Cycle. .......................... 22
5. The Completed 25-Year Cycle .. ................................ 25
6. Comparison of Schematic and Observational Years. ............. ............. 28
7. Names of the Months .. ... ............. ................... 45
8. Chronology of the 12th Dynasty. .. ............ ................... 69
Xi
oi.uchicago.edu
oi.uchicago.edu
LIST OF ABBREVIATIONS
AJSL American Journal of Semitic Languages and Literatures. . . Chicago, etc.,
1884-1941. 58 v.
Ann. Serv. Egypt. Service des antiquites. Annales. Le Caire, 1900-
BASOR American Schools of Oriental Research. Bulletin. South Hadley, Mass., 1919-.
JAOS American Oriental Society. Journal. Boston, etc., 1849-.
JNES Journal of Near Eastern Studies. Chicago, 1942-.
LD Lepsius, Richard. Denkmaeler aus Aegypten und Aethiopien . . . Berlin, 1849-56;
Leipzig, 1897-1913. 19 v.
MH Chicago. University. Oriental Institute. Epigraphic Survey. Medinet Habu I-.
Chicago, 1930-.
OLZ Orientalistische Literaturzeitung. Berlin, 1898-1908; Leipzig, 1909-.
PSBA Society of Biblical Archaeology, London. Proceedings. London, 1879-1918. 40 v.
Urk. Urkunden des aegyptischen altertums. . . . Leipzig, 1903-.
Wb. Erman, Adolf. Worterbuch der aegyptischen Sprache, im Auftrage der deutschen
Akademien hrsg. von Adolf Erman und Hermann Grapow. Leipzig, 1926-31. 5 v.
Same. Die Belegstellen. 1. Heft. Leipzig, 1935. 2. Bd., Heft 1-. Leipzig,
1937-.
ZAS Zeitschrift fuir agyptische Sprache und Altertumskunde. .. . Leipzig, 1863-.
xiii
oi.uchicago.edu
oi.uchicago.edu
INTRODUCTION
1. Since the astronomical problems with which we shall deal in this book are almost entirely lunar, I shall
describe as simply as I can the more important aspects of lunar motion. Before we turn to that, however, a cer-
tain amount of groundwork is necessary.
2. THE CELESTIAL SPHERE. In Figure 1 we have the earth shown at the center of a much larger sphere,
upon which we may assume are located all the stars and planets without regard to their actual varying distances
from the earth. We shall call this the celestial sphere. We know that the earth itself is a sphere rotating around
an axis with north (N) and south (S) poles. At right angles to this axis is a great circle (WE) whose plane passes
through the center of the earth. This circle we call the equator.
3. By extending to the celestial sphere the axis and the equator of the earth, we have celestial poles (N',S')
and a celestial equator (W'E'). It is true that our earth moves around the sun; but the sun is the only star around
which the earth revolves, and all the others are so vastly distant that the movement of the earth has no appreci-
able effect on the location of the celestial poles or the celestial equator.
FIG. 1.-The celestial sphere with the earth at its center.
4. THE ECLIPTIC. Owing to the rotation of the earth on its axis, in the daytime the sun rises in the east
and sets in the west. At night the stars, or at least some of them, rise in the east and set in the west. Some
stars, to be sure, never rise and set but simply revolve about the pole. Now, if the sun were dimmed sufficiently,
we should see it and the stars in the daytime rise and set; but, because the earth goes around the sun, it would
be seen, from the earth, against a slightly different background of stars from day to day. Figure 2 is the celes-
tial sphere seen from above, with the sun at the center and the earth in its orbit. When the earth is at a, the sun
oi.uchicago.edu
INTRODUCTION
would be seen against the celestial sphere at a'. One day later the earth would have moved to b, and the sun
would be seen at b'. After about 91 days, the earth would have moved to c, one-quarter around its orbit, and
the sun would be seen at c'. In a year the earth would have moved entirely around the sun, and the sun in turn
would have apparently moved in a complete circle among the stars. This apparent path of the sun is called the
ecliptic.
C
b
FIG. 2.-The celestial sphere with the sun at its center and the earth
in its orbit around the sun, to demonstrate the ecliptic.
5. THE VERNAL POINT. Both the ecliptic and the equator are great circles on the celestial sphere with
the earth at their center. These two circles are shown in Figure 3. Corresponding to the tilt of the earth's axis
of 23 1/20, the plane of the ecliptic makes an angle of 23 1/20 with the plane of the equator. One half of the eclip-
tic belongs to the Northern Hemisphere, one half to the Southern Hemisphere. The point where the ecliptic, go-
ing from south to north of the equator, intersects it is called the vernal point (T) or the spring equinox. It is
from this point that the zodiac and the positions of sun and moon (longitude) on the ecliptic are measured.
6. THE MOON. The path of the moon around the earth varies but is nriever farther than 50 8' from the eclip-
tic. While the sun apparently goes round the ecliptic but once a year, the moon actually goes round it twelve and
a fraction times; and every time it does so we have a lunar month.
7. LUNAR MONTHS. When the sun and moon are at the same point on the ecliptic (in other words, when the
sun, moon, and earth are in a line), we have the moment called conjunction. The moon moves around the ecliptic
and back to its starting point in from 27.18 to 27.47. on the average 27 1/3, days, a sidereal month. In that time,
however, the sun itself has moved, and it takes the moon about two days more to catch up with it again for con-
junction. The time from conjunction to conjunction (or from full moon to full moon) is called a synodic month,
and it averages 29.53059 days. Lunar calendars are always based on synodic months.
8. Another way of explaining these two terms is illustrated in Figure 4. In A the sun, moon, and earth are
in line with a star on the celestial sphere, and the moon is at point c in its orbit (here shown as a simple circle).
In B the moon has gone completely around the earth and has reached c again, in line with the same star. This is
oi.uchicago.edu
FIG. 3.-The vernal point, where the ecliptic, going from south to north, crosses the equator.
Sun
A B
FIG. 4.-The sidereal and the synodic months.
3
oi.uchicago.edu
INTRODUCTION
the sidereal month. But, since the earth itself has moved during these days along its orbit, the moon is obliged to
travel about two days more before it is again in line with the sun. When it is again in line, a synodic month has
elapsed.
9. THE SYNODIC MONTH. It has been stated that the synodic month averages 29.53059 days. Some months
may be longer than that and some shorter. Like the orbit of the earth about the sun, the orbit of the moon is an el-
lipse. When it is closest to the earth (perigee), it is 221,463 miles away. When it is farthest (apogee), it is distant
252,710 miles. When the earth is farthest from the sun (aphelion) and the moon at conjunction is at perigee, the
moon will be traveling faster (since every satellite travels fastest when nearest the primary body) and will have
less distance to travel from the point c (where it has completed a sidereal month) to conjunction than when the earth
is nearest the sun (perihelion) and the moon is at apogee. In Figure 5, A illustrates the first situation and B the
second. The shortest possible synodic month is 29.26 days and the longest 29.80 days.
Sun Sun
Apogee
Perigee
Perihelion
Aphelion
A B
FIG. 5.-The variation in the length of the synodic month.
10. THE PHASES OF THE MOON. The moon shines by light reflected from the sun. Conjunction (when the sun,
moon, and earth are in line) is a purely astronomical event, not visible to man (except, of course, when it causes
an eclipse). It may occur at any hour of the day or night. In 24 hours after conjunction the moon will have moved
on the ecliptic slightly more than 120 away from the sun. This may be sufficient for it to be visible as a thin cres-
cent in the western sky after sunset. When the moon has moved 90

from the sun, it is at first quarter. At 180

it
is directly opposite the sun, and we have full moon. At 2700 we have third quarter, with the moon visible only in
the latter part of the night. Then it draws closer and closer to the sun, until one morning just before sunrise it is
visible, again as a thin crescent, for the last time before conjunction.
11. NEW CRESCENT VISIBILITY. Since conjunction is invisible, the lunar month began for most primitive
people with the reappearance of the moon as a crescent. The time that must elapse after conjunction for visibility
to be possible is variable. At Babylon (lat. 32.50), it varies from a minimum of 16.5 hours to a maximum of about
42.1 The factors which control this are three: the anomaly of the moon (its distance from the earth), the obliquity
oi.uchicago.edu
INTRODUCTION
of the ecliptic (its angle with the celestial equator), and the latitude of the moon (its distance north or south of
the ecliptic). We shall consider each in turn.
12. Figure 6 illustrates the effect of the anomaly of the moon on crescent visibility. After conjunction, in
both diagrams the moon travels the distance to c in 24 hours. At apogee the moon travels less rapidly than at
perigee, and this, combined with its greater distance from the earth, results in a smaller angle between sun,
earth, and the moon than at perigee. Consequently, at perigee the moon will be relatively farther away from the
sun at sunset and more likely to be visible.
Apogee
Sun Perigee
Earth
Earth
FIG. 6.-The effect of the anomaly of the moon on crescent visibility.
13. In Figure 7 the effect of the obliquity of the ecliptic is diagrammed. In March, at the vernal equinox,
the sun moves to the north of the celestial equator on the ecliptic, which has a higher angle with the horizon by
23 1/20 than the equator has. In diagram A the sun and moon are so close together that no visibility is possible.
One day later, in B, the moon has dropped back another 120 (about 50 1/2 minutes) from the sun, and the result-
ing effective distance (a-b) is great enough for visibility.
A
March
B
C
5eptember
D
FIG. 7.-The effect of the obliquity of the ecliptic on crescent visibility.
oi.uchicago.edu
INTRODUCTION
14. In C we have the September counterpart of A. The sun is now moving to the south of the celestial equator
on the ecliptic, which is now 23 1/2* lower toward the horizon than the equator. The moment is just after conjunc-
tion, and no visibility is possible. In D , one day later, the moon has dropped back the same distance as in B
along the ecliptic, but its effective distance from the sun (a-b) may still be too small to permit visibility.
15. The effect of the latitude of the moon is pictured in Figure 8. Diagram A is based on Figure 7 B and shows
the possibilities of the moon being 5

north or south of the ecliptic on the effective distance between sun and moon
in March. B reflects the situation in Figure 7 D, again with the moon both north and south of the ecliptic. Its
bearing on the effective distance between sun and moon in September is obviously much greater. Schoch com-
ments: "In spring the mean anomaly of the Moon is of the same importance as her latitude, but in autumn her
latitude is far more important."
2
Sun
e .g
March September
A B
FIG. 8.-The effect of the latitude of the moon on crescent visibility.
16. One factor, independent of the motion of the moon, which also affects crescent visibility is the latitude of
the observer. The farther north one goes, the lower in the sky will appear the celestial equator, and the lower,
both in March and in September, will be the angle of the ecliptic with the horizon and the smaller will be the dis-
tance a-b. I have given above the minimum and maximum hours required for visibility at Babylon as stated by
Schoch. He also gives the figures for a place in latitude 51N., and they are 20 hours minimum and 63 hours maxi-
3
mum.
17. OLD CRESCENT VISIBILITY. All the factors which combine to produce the great variability in time re-
quired after conjunction for visibility of the new crescent also affect the last visibility of the old crescent. Gen-
erally speaking, the time from last possible visibility of the old crescent to conjunction is the same as the time
required after conjunction for first visibility of the new crescent. It may be a little more or less, depending upon
the trend of the factors affecting visibility.
18. SEQUENCE OF LUNAR MONTHS. If the length of the synodic month were exactly 29.5 days, and the time
required for crescent visibility were a constant, we should have a regular sxuccession of lunar months with first
29 days and next 30, then 29 again, and so forth. Since, however, the length of the synodic month varies and the
time required for crescent visibility also varies, it is quite possible to have two 30-day months or two 29-day
months in a row. When the synodic month is below average length and the time required for visibility is small, it
is possible to have three 29-day months in a row. Conversely, when the synodic month is lengthening beyond the
average and the time required for visibility is also lengthening, it is possible to have three, at times four, and
very rarely five 30-day months in a row.
19. FULL MOON. The time required from conjunction to full moon is also variable. In Figure 9 A the moon
travels through perigee from conjunction to full moon, and the time is shorter because the distance is less and its
speed is faster than in B, where it travels through apogee. The necessary time for full moon varies from 13.73
to 15.80 days after conjunction.
oi.uchicago.edu
INTRODUCTION
Sun
Monat
Full Moon
0-Conjunction
Earth
Perigee
A
Sun
SMoon
atFull Moon
Conjunction
Earth
B Apogee
FIG. 9.-The variability of time between conjunction and full moon.
20. TABLES FOR CALCULATING THE MOON. The best and easiest tables to use in calculating conjunction
are those of Carl Schoch in Langdon and Fotheringham, The Venus Tablets of Ammizaduga (London, 1928). They
are for Babylon, but a correction for difference in longitude is easily made. His Table G may be used for an ap-
proximation of the time from conjunction to new crescent or from old crescent to conjunction. Since all Egypt lies
south of Babylon, the time required in either case will not be greater than the figure derivable from Table G. Any
doubtful case may be checked by the use of tables E 21-26 in P. V. Neugebauer, Astronomische Chronologie (Berlin
and Leipzig, 1929), both for old and for new crescent. The elements 0 (the longitude of the sun), X (the longitude
of the moon), and ,p (the latitude of the moon), which are necessary for the calculation, may be conveniently derived
from Carl Schoch's Planetentafeln fir Jederman (Berlin, 1927) or from P. V. Neugebauer's Tafeln zur astronomi-
schen Chronologie (Leipzig, 1912-25).
21. HELIACAL RISING OF SIRIUS. Aside from the moon, we shall be concerned in this book only with the star
most important to the Egyptians, Sirius, or Sothis, as the Greeks rendered its Egyptian name. The stars which
rise and set have, like the moon, conjunctions with the sun, and, like the moon, they are invisible when near con-
junction. This is not, of course, because they shine by reflected light, but because the sun is so much brighter than
any star that it simply swallows up the star's light. Unlike the moon, each star has but one conjunction a year, and
its period of invisibility is much greater than the moon's. That is because the sun apparently moves among the
stars at a rate of about 10 a day (about four minutes) and various other factors, such as distance from the ecliptic,
also have a bearing. Sirius is invisible for a period of 70 days. There comes a time, however, when the sun has
dropped back from it enough so that it is again visible in the dawn, just before sunrise, above the eastern horizon.
This event is called its heliacal rising, and the factors that govern it are the arcus visionis (f3), that is, the height
of the star above the horizon which is necessary for visibility, and the latitude of the observer (q'). In Egypt, gen-
erally speaking, an advance southward by 10 of latitude means a heliacal rising of Sirius earlier by one day. The
experimental data we now have for 1 , though limited, give it a range from 9.4o to 8.60.5 For Heliopolis (4 = 30.10)
and Memphis (4= 29.9) these values result in a heliacal rising on July 17-19 (Julian) throughout Egyptian dynastic
history.
22. THE CIVIL CALENDAR. In the following pages when the term civil calendar (abbreviated "civ.") is used, it
will mean the familiar Egyptian year of 365 days, comprised of three seasons, with four months each, 30 days to
each month, with the 5 extra days called epagomenal after the twelfth month. Dates in this calendar will be given
in the form II prt 23 (meaning the second month of the season prt, day 23). In the New Kingdom and later the
months were sometimes referred to by names (226-30 and Table 7). These are:
oi.uchicago.edu
First Season
Tht
1. Thoth
2. Phaophi
3. Athyr
4. Choiak
INTRODUCTION
Second Season
prt
5. Tybi
6. Mechir
7. Phamenoth
8. Pharmuthi
Third Season
mw
9. Pachons
10. Payni
11. Epiphi
12. Mesore
23. OTHER CALENDARS. The Alexandrian calendar (abbreviated "Al.") is simply the civil calendar arrested
in its forward shift by the addition of a sixth epagomenal day every fourth year. Whether this calendar was insti-
6
tuted in 30 or in 26 B.C. is still in dispute;6 but in its actual working Thoth 1 (I .t 1) Al. begins on August 29 (in
leap year August 30), and this was the situation obtaining in 26-23 B.C.
24. In 46 B.C. Julius Caesar introduced the Julian calendar, which is the same as our modern Gregorian cal-
endar except for the fact that no leap years are ever passed over. The Julian calendar, projected backwards,
is the one customarily used to express dates in ancient history and for astronomical calculations. When, for the
sake of the exact season of the year, it is desired to convert a Julian date into the corresponding Gregorian date,
recourse may be had to Table 1.
CONVERSION
TABLE 1
OF JULIAN INTO GREGORIAN DATES
From To
Mar. 1 Feb. 29
701 B.C. 601
601 501
501 301
301 201
201 101
101 A.D. 100
A.D. 100 200
200 300
300 500
500 600
600 700
700 900
900 1000
1000 1100
1100 1300
1300 1400
1400 1500
1500 1700
1700 1800
1800 1900
1900 2100
Subtract
in Days
7
6
5
4
3
2
1
0
Add in Days
1
2
3
4
5
6
7
8
9
10
11
12
13
From
Mar. 1
3701 B.C.
3501
3401
3301
3101
3001
2901
2701
2601
2501
2301
2201
2101
1901
1801
1701
1501
1401
1301
1101
1001
901
To
Feb. 29
3501
3401
3301
3101
3001
2901
2701
2601
2501
2301
2201
2101
1901
1801
1701
1501
1401
1301
1101
1001
901
701
Subtract
in Days
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
oi.uchicago.edu
CHAPTER I
THE BEGINNING OF THE EGYPTIAN LUNAR MONTH
25. THE PROBLEM. All that we know of ancient and modern time-reckoning leads to the conclusion that
lunar months begin with some observable phase of the moon. UAs always," says Nilsson, "the concrete phenome-
non is the starting point."
'
Most peoples (both ancient and modern) who use a lunar calendar start the month
with the new crescent;2 a few count from full moon;3 while two East African tribes, the Masais and the Wadschag-
gas, begin with the moon's invisibility.4 All these starting points, as well as one other, conjunction, have been
maintained for the Egyptian lunar month by various scholars from as recent a year as 1932. It is the purpose of
this chapter to demonstrate that the ancient Egyptians, like the Masais and the Wadschaggas, began their lunar
month on the morning of the day when the old crescent of the moon was no longer visible in the eastern sky before
sunrise.
26. EARLIER VIEWS. Prior to 1864, when anyone spoke of an Egyptian lunar month, he undoubtedly meant
one beginning with the new crescent. Lepsius, the first chronologer to have available much original source mate-
rial for a study of the Egyptian calendar, came to the conclusion (in 1849) that "Fragen wir endlich noch nach dem
Anfange des Mondjahres, so kann es an sich nicht zweifelhaft sein, dass es mit einem Neumonde begann."
5
Be-
fore Champollion all calendar study was dependent on classical sources, often vague and contradictory; and,
since both Greeks and Romans had lunar calendars based on visibility, it is nowise strange that in the pages of
Ideler, who brilliantly summed up the Egyptian calendar in 1825, no suggestion can be found of an Egyptian lunar
month beginning at any other time than new crescent.
6
27. In 1864 Brugsch published a text from the propylon of the temple of Khonsu in Karnak (quoted below in 38)
which led him to the conclusion that the month started with conjunction, that the crescent appeared on the 2d day
and full moon on the 15th.
7
In this opinion he was followed by later scholars such as E. Mahler and Sethe. The
obvious and weighty objection to conjunction as the starting point is, of course, that this event is not observable
and is not, therefore, one of the concrete phenomena spoken of above. No people, unless indeed it be the Egyptians,
ever began the month with conjunction. Sethe recognized the difficulty,10 but Mahler was the only one who at-
tempted to meet it. He suggested that the Egyptians determined by observation the approximate time of the full
moon and from that point, since they were well aware of the average length of the synodic month, they counted
about 14 1/2 days and arrived at the time (if not to the hour and the minute, still certainly to the day) of conjunc-
tion.11 This hypothesis of Mahler's completely lacks supporting evidence and can hardly be said to meet the seri-
ous criticism one can easily bring against it. We have seen, for example, that the full moon varies from 13.73 to
15.80 days after conjunction (19). If we add 14.5 to both extremes, we should have synodic months of 28.23 to
30.30 days in length, whereas we know that they vary only from 29.26 to 29.80 days (9). Mahler's method would
at times inevitably lead to months of either 28 or 31 days, unless some adjustment were made, and this he does
not suggest. Furthermore, as we shall see, the Egyptians counted the 15th day as that of full moon, without re-
gard to astronomical exactness.
28. These, and possibly other considerations, induced D. R. Fotheringham,12 Lehmann-Haupt,13 Meyer,14
and more recently Edgerton
5
and Wood
6
to maintain crescent visibility as the logical starting point of the
month, though none, except Lehmann-Haupt in slight degree, has argued the problem. Standing quite alone is
Macnaughton in proposing that the lunar month began originally with full moon, though he admits the possibility
that, in the 18th dynasty and later, it may have begun, like that of the Babylonians, with new crescent.
1 7
Since
the extreme unlikelihood of this proposal's being correct will be brought out in the following pages, it here re-
quires no further consideration.
29. We come now to the somewhat contradictory, or at least incompletely expressed, view of Ludwig Bor-
chardt. In 1925, reviewing Nilsson, Primitive Time Reckoning, he stated: "Ferner, wenn wir aus den jiberliefer-
ten Namen der Mondmonatstage nachweisen k innen, dass die alten Agypter ihre Mondmonate . . . mit dem ers-
oi.uchicago.edu
THE CALENDARS OF ANCIENT EGYPT
ten Tage, an dem sie den Mond nicht mehr sahen, also mit dem Tage nach dem Altlicht, begonnen ... , so wird
man Ahnliches vergeblich bei anderen alten Kulturvolkern suchen.
"
18
30. This statement is clear enough but is nowhere elaborated. Instead, in the pages of Die Mittel zur zeit-
lichen Festlegung von Punkten der agyptischen Geschichte und ihre Anwendung (Kairo, 1935)19 there are scat-
tered remarks which only confuse the reader. In one place (p. 30, n. 1) he reaffirms the day after "Altlicht," old
crescent, to be the month's beginning, but then (p. 6, n. 1; p. 30, n. 1) he mentions "Altlicht" itself as the starting
point, and still elsewhere (pp. 7, 24, 25, 36, and 37) he speaks of "wahrer Neumond," conjunction, as though that
ought to be the first day of the month. The whole matter is further complicated by his belief that at an early date
the Egyptians adopted a conventionalized lunar calendar of alternating 29- and 30-day months, with now and again
one of 31 days.
2 0
The result is, "dass der erste kalendarische Monatstag nicht genau auf den Tag des astronmi-
schen Neumonds zu fallen braucht."
2 1
His last published remark on the subject speaks of "unseren geringen
Kenntnissen vom Aufbau des igyptischen Mondjahres" and suggests that a certain lunar date he is discussing may
be either (1) a "kalendarischer Neumond" or (2) an "Altlicht-Tag," which can be the beginning of a lunar month
"nach der alten Art," or (3) the day before "Altlicht," in which case it could denote the end of a lunar month.
2 2
31. It is probably safe to conclude, despite these contradictions, that Borchardt would have the earlier Egyp-
tians begin their lunar month after "Altlicht." So far as his publications indicate, his evidence for this theory is
(1) the fact that in the Medinet Habu calendar the recurrent monthly lunar feasts are listed in the order (by the
day) 29, 30, 1, 2, 4, 6, 10, and 1523 and (2) his conviction that the third day of the month was named after and
marked by the appearance of the crescent.
2 4
This last is, as we shall see (38), in contradiction to the statements
of the Egyptians themselves that the crescent came on the second day of the month, and the former, to my mind,
represents simply a convenient grouping of feasts by proximity. The one or two feasts which occur immediately
before the first day of the month are listed there, rather than after the feast on the 15th, with which they have no
temporal connection. It is perfectly true, as I hope to demonstrate, that "Altlicht" fell on either the 29th or the
30th, so that the list does, in Borchardt's words, "eine Reihe von Altlicht bis Vollmond bilden";
2 5
but the truth
of this statement can be apparent only after it has been demonstrated on other grounds that the month began with
the day after old crescent.
We now turn to a consideration of the data which, I believe, will conclusively prove this thesis.
32. THE BEGINNING OF THE DAY. That the day in Egypt began at dawn, and was reckoned from one dawn to
the next, has been fully demonstrated.
2 6
Since the evidence seems conclusive to me, I shall not detail it again
but shall go on to the bearing of that fact on the lunar month. It is obvious that, when the month begins, the 1st
day of the month also begins. Those peoples who begin their day in the evening do so because they base their
month on visibility of the new crescent or on full moon, events which are observable in the evening.27 If the be-
ginning of the Egyptian day is connected with the lunar month, then we must seek a lunar phenomenon associated
with the morning. This can only be the gradual waning of the moon in the eastern sky until it is just visible be-
fore sunrise on one morning (old crescent), while on the following morning it is invisible.
33. Sethe has a different explanation. According to him, morning beginning is the necessary consequence of
the regulation of the year. "Da das Normaljahr mit dem Friihaufgang des Sirius begann, der eine Stunde vor Son-
nenaufgang sichtbar wurde, so musste auch der Tagesbeginn auf den Morgen fallen.",
2 8
I shall argue later (187-
211) that no such year as a "Normaljahr" existed, and I do not wish to anticipate that discussion here. But Sethe,
in company with many others, believed that a lunar calendar or, at least, the lunar month was used in Egypt be-
fore the civil year was inaugurated,
2 9
and it is universally accepted that as a measure of time the day and the
lunar month came before the concept of the year.
34. If the Egyptians had first of all a lunar calendar, then their day began when their lunar month began. The
beginning of their day then either remained the same throughout their history or it moved from evening to morn-
ing. No explanation of the origin of the civil calendar yet given would require such a shift. We may reasonably
conclude that it never occurred and that the beginning of the day in Egypt is inseparably connected with the lunar
month.
35. I do not wish to push this argument too strongly. If it stood alone, it would be far from conclusive; but
it does fit in nicely with the other evidence to be presented, and it can certainly be regarded as confirmatory.
10
oi.uchicago.edu
36. THE NAMES OF THE DAYS OF THE MONTH. Comparative lists of the names given to the days of the
30 31
lunar month can be found in Brugsch, Thesaurus, pages 46-48. Table 2 is eclectic and does not present all
possible variations. A late list from the temple of Edfu appears on Plate V. The interpretations given to cer-
tain names I shall endeavor to establish in later pages.
TABLE 2
THE DAYS OF THE LUNAR MONTH
1. e psdtyw; later vars. and 0 psdntyw*
2. ~*' *' * ~ (tp) :bd, "new crescent day."
3. __K 7mspr, (arrival' day."
4. prt s'm, "day of the going-forth of the sm-priest."
5. ,1A' ' ibt Ihr 13wt, "'day of offerings on the altar."
6. 4A%% snt, "6th day."
7. C dnit, "part day, first-quarter day."
8. tp.
9. i}( k p
1. sifon ;f
11. s tt.
12. reading uncertain.
13. o m ' m sy.
14. Z~Z26w
15. - 7 I tp) s'mdt, "half-month day, 15th day, day of full moon."
16. mspr sn-nw, "second rarrival' day."
18. 9' 7 ji~1, "day of the moon."
24. _T7 knlw.
25.s tt.
26. L]7 prt.
11
oi.uchicago.edu
TABLE 2 -Continued
27. J wb.
28. hb-sd Nwt, "day of the jubilee of Nut."
29. h ....
nz
30. OA/T7 prt Mn, "day of the going-forth of Min."
* The use of hieroglyphic type to reproduce most of the texts quoted in this study requires a word of caution.
Some signs in the late texts are reversed, especially the rnpt sign, and unimportant details both in form and in
position are disregarded.
t WB., I, 65, reads ibd, but see A. H. Gardiner, Egyptian Grammar (Oxford, 1927), p. 475, N 11.
$ Langdon, Babylonian Menologies and the Semitic Calendars (London, 1935), p. 91, n. 6, believes that the
Egyptian root s-m-d is an Egyptian form of the root found in Babylonian as tabatu, "to divide, cut off, come to
an end," Arabic sabata. Sapattu or Sabattu, the 15th day of the Babylonian month, actually means "the end of
the first half of the month" (op. cit., pp. 90-92). Gardiner, Grammar, p. 475, N 13, rejects the reading smdt
entirely, quoting such a writing as o . This seems to me merely a conflate form such as we have later
in ni mQ7. See my further remarks on this in 42, below.
37. Some of the names, it is apparent, are phases of the moon, others are numerical, while yet others are
concerned with ritual or mythology. Many are untranslatable, at least with any certainty.
38. The three days concerned most importantly with the phases of the moon are brought out in the following
text from the Ptolemaic propylon of the temple of Khonsu in Karnak:
3 2
"He (Khonsu, the moon-god) is conceived (bk-) on psdntyw; he is born on bd; he grows old after smdt'." The
ready paraphrase is that he is conceived in the darkness of invisibility on the first day of the month, that he is
born as the new crescent on the second day, and that he wanes after the day of full moon, the 15th day.
39. An earlier text from the Middle Kingdom confirms this:
3 3
"I know, O souls of Hermopolis, what is small on the 2d day and what is great on the 15th day; it is Thoth."
Thoth is, of course, the moon, small on the day of new crescent and great on the day of full moon.
40. That the 15th day was that of full moon is substantiated by many other texts, some of which have been
dealt with by Brugsch.
3 4
I shall quote only one more text, from the temple of Edfu:
3 5
"He created the 15th day to illuminate (hd)the land throughout the night."
41. In the light of these texts, the identification of gbd and ~mdt as "new crescent day" and "full moon day" can
hardly be questioned. The precise meaning of the name of the first day pid(n)tyw, remains uncertain. Since in
36
the Pyramid Texts it is written without the n, it is possible to see it as "Ennead day," or the like. Borchardt
believed that it could have nothing to do with the pid of either "nine" or "shine," but preferred to see in it a ref-
erence to "back," with the suggestion that the Egyptians may have seen the moon, at conjunction, as turning its
back, its dark side, to earth.3
7
Sethe, on the other hand, saw a possible connection between "nine" and "new-
(month)," citing the correspondence, in the Indo-Germanic languages, between the roots for "neun" and "neu.,
42. I can add nothing to the discussion, except a suggestion that the n appearing in a later variation of the
i- n
name is there through analogy with such writings of the 6th and 15th days as ,m which have
the element nt added to the numeral. Neither in the Pyramid Texts nor in the Old Kingdom generally is
psd(n)tyw written with the n. It seems to appear first with n as in the list of feasts on the Middle King-
dom coffin of Ma ( ).40 Three times in the Middle Kingdom
4 1
and at least once later (in Pap. Boulaq 17, v, 2,
12
oi.uchicago.edu
the Cairo hymn to Amon) the numeral is used for psd. According to Borchardt's description, the Middle Kingdom
examples are written thus I , ; the later one is . If the pd of psdtyw either was, or was thought to
be, the numeral "nine," the n might easily have been introduced by analogy.
43. It is time now to leave philology and examine the purely astronomical basis underlying the choice of names
for the days of the lunar month. By this means we shall be able easily to see the significance of the names of cer-
tain days which have not yet figured in the discussion.
44. THE ASTRONOMICAL BASIS FOR THE NAMES OF THE DAYS OF THE MONTH. In the latter part
of this chapter I have had occasion to make sixty-five calculations of conjunctions with their accompanying morn-
ings of crescent invisibility and evenings of new crescent visibility. In forty-six cases (70 per cent) the crescent
was visible on the evening of the day (Tbd) after that on which there was no lunar visibility in the morning (psdntyw).
In the other nineteen cases (30 per cent), new crescent was first visible on the third day of the month, mspr, which
I have termed "'arrival' day."
4 2
45. In Figure 10 is diagrammed the astronomical situation at the beginning of the month. Just before dawn on
either the 29th or the 30th day, the last crescent is still to be seen. On the following morning it cannot be seen,
and the new month begins with psdntyw. Just after sunset on the following day, ibd, the new crescent is visible in
seven months out of ten. In the other three months it is first seen on mspr.
46. When new crescent can be seen on -bd, there is a period of 60 hours from old crescent to new. The mean
time of conjunction is thus at noon on psdntyw. If we take 17 hours as the minimum time that must elapse from
last visibility to conjunction and from conjunction to new visibility (11-17), we have a possible range of time,
during which conjunction may occur, of about 26 hours.
4 3
When the number of hours required for visibility in-
creases to 30, or the hours required plus the number of hours after noon on psdntyw to time of conjunction total
30 or more, then new crescent is delayed until mspr. In that event the time from last to new crescent is 84 hours,
mean conjunction is at midnight on psdntyw, and the possible range is as shown.
4 4
It is instructive to note that
most conjunctions fall on psdntyw (so in fifty-seven out of my sixty-five calculations), a few on the day before
(seven out of sixty-five), and still less (one out of sixty-five) on the following day. Thus Brugsch, Mahler, and
Sethe were roughly correct in their theory (27). Their error was in failing to associate psdntyw with an observ-
able phenomenon.
47. On the basis of Figure 10 and the time required from conjunction to full moon (13.73-15.80 days; cf. 19),
it is possible to diagram the situation for that time of month. Figure 11 shows the possibilities. When new cres -
cent is visible on 8bd, then mean full moon is just at the end of smdt with a possible range before and after of
some 72 hours. When the new crescent is delayed until mspr, then mean full moon is also delayed to the begin-
ning of the night on mspr sn-nw, "second rarrival' day," to my mind a deliberate and meaningful choice of name.
In no case does full moon ever occur earlier than the night of the 14th or later than the night of the 17th;45 both
these days bear the same name, 9l3w, and, while I am unable to translate this, I cannot believe that it is lacking
in significance.
48. Interesting relationships exist between other days of the lunar month.46 As we should expect, the day of
full moon is exactly between day 7, dnit, "first quarter," and day 23, dnit, "last quarter." The day after the full
moon group (Fig. 11) is day 18, ih, "day of the moon," possibly to be regarded as the definite beginning of the
second part of the month. From crescent to full moon there are 13 days. Thirteen days after day 18, assuming
a full month of 30 days, a new month begins. Day 11, seven days before day 18, bears the same name, stt, as does
day 25, seven days after day 18. This last seems to indicate that some importance was attached to day 18, but
other than that its meaning is unclear to me.
49. THE LUNAR CALENDAR OF PAPYRUS CARLSBERG 9. There are a few events in Greco-Roman
times dated both to the civil calendar and to the day of the lunar month. These are of inestimable value in check-
ing the results attained above, but they may best be considered after we have first examined the only truly mathe-
matical astronomical Egyptian text yet published, Papyrus Carlsberg 9,
4 7
whose importance for our purpose can
hardly be overstated.
50. This papyrus was written in or after A.D. 144, and it furnishes a simple scheme, based upon the civil
calendar, for determining the beginning of certain lunar months over a 25-year cycle. Underlying this cycle are
13
oi.uchicago.edu
Crescent Visible onl 3rd Day
H - --------------
42 hours
-----------
-----------
- -4
r s e t V sb e o
n
a
- - - - 0h u s--------------
us
---
42--
I I -sdntyw/ II b
DAY 29 OR 30 I / ibd
I - I I 2
NIGHT
DAY
LAST VISIBLE CRESCENT
FIRST VISIBLE CRESCENT
HOURS REQUIRED FOR CRESCENT VISIBILITY
HOURS AFTER NOON OF psd nty Lt TO CONJUNCTION
RANGE IOF CONJ./~I '
R=17 I /
OR R+ T= LESS THAN 30 I
I 7 I
RANGE OF CONJ.
R=30
OR R+T = 30 OR MORE
FIG. 1.-The beginning of the lunar month.
hours-------- ---- -
mspr
3
Range of Full Moon
Crescent Visible 3rd Day
K----- 36 hours---------- ----------- 36 hours-- -----
I Range of Full Moon
I Crescent Visible 2nd Day l
-- -36 hours ------- ------ - 36 hours -----------
mspr sn-nw SS3W
'7 18
- NIGHT
DAY
Q MEAN FULL MOON
FIG. 11.--The possible range of full moon.
R
T
f-+
9W
'4
.gmd t
15
i
0
oi.uchicago.edu
the facts that 25 Egyptian civil years have 9,125 days and that 309 lunar months (divided into 16 years of 12
months and 9 years of 13 months)48 have 9,124.95231 days. The earliest cycle mentioned in the text began in
the sixth year of Tiberius, A.D. 19, and the latest in year 7 of Antoninus, A.D. 144. Actually, the calendar does
not indicate when every month began but lists only six dates for every year, those falling in the second and fourth
months of each season. It also indicates the years of 13 months ("great" years)49 according to the following
scheme: 1st, 3d, 6th, 9th, 12th, 14th, 17th, 20th, and 23d year of each cycle.
THE 25-YEAR CYCLE
Months: I
Year 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
II
1
20
9
28
18
7
26
15
4
24
13
2
21
10
30
19
8
27
16
6
25
14
3
22
12
III
TABLE 3
AS GIVEN IN PAP.
I II
29
18
7
26
16
5
24
13
2
22
11
30
19
8
28
17
6
25
14
4
23
12
IIII
30
19
8
27
17
6
25
14
3
23
12
1
20
9
29
18
7
26
15
5
24
13
2
21
11
20
10
PRT
III IIII
28
17
6
25
15
4
23
12
1
21
10
29
18
7
27
16
5
24
13
3
22
11
30
19
9
CARLSBERG 9
SMW
I II III IIII
27 26
16 15
5 4
24 23
14 13
3 2
22 21
11 10
30 29
20 19
9 8
28 27
17 16
6 5
26 25
15 14
4 3
23 22
12 11
2 1
21 20
10 9
29 28
18 17
8 7
51. Table 3 gives the complete cycle as stated in the papyrus. Anyone using this calendar would begin a
lunar month, without regard for actual observation, on II 1t 1 (which for A.D. 19 would be September 18; for
A.D. 44, September 11; for A.D. 69, September 5; and so on up to A.D. 144, August 17), on IIII t 30 (for A.D. 19,
December 16; for A.D. 144, November 14), on II prt 29 (for A.D. 20, February 13; for A.D. 145, January 12), and
so forth.50 What is significant about the cycle for our purpose is that it gives a continually recurring series of
lunar dates, easily and accurately translatable into the Julian calendar for this period.51 That means it should
also be possible, by calculating the lunar phenomena fitting any series of dates, to determine what is the govern-
ing principle of the beginning of the month.
52. The editors of the papyrus saw this, and, after studying the dates for the year A.D. 144, they came to the
conclusion that the month began with new crescent visibility.52 Now if we calculate conjunction, morning of in-
visibility, and evening of visibility (cf. 20) for the first two years, A.D. 144 and 145, of a cycle, we arrive at the
following
results:
5 3
15
oi.uchicago.edu
53. First Two Years of Cycle Beginning A.D. 144
1st Day of
Lunar Month
Aug. 17, 144
Nov. 14
Jan. 12, 145
Mar. 12
May 10
July 8
Sep. 5
Nov, 3
Jan. 1, 146
Mar. 1
Apr. 29
June 27
Conjunction
Aug. 16, 6:31
Nov. 13, 12:13
Jan. 11, 1:01
Mar. 11, 7:07
May 9, 3:07
July 7, 6:55
Sep. 4, 6:25
Nov. 2, 3:52
Dec. 31, 1:15
Feb. 28, 12:54
Apr. 28, 4:01
June 26, 9:35
54. In nine out of twelve cases, the lunar month begins on the same day as crescent visibility, and this cer-
tainly seems to bear out the editors' conclusion. Does that finding invalidate all our earlier argument? We shall
see that it does not, Let us recall that 309 lunar months have 9,124.95231 days while 25 civil years have 9,125
days. Obviously the lunar calendar has a lag of .04769 day every 25 years. In 500 years this would be roughly
54
one day.54 Let us test this by calculating, as we did above, the first two years of the cycle beginning 1,000 years
before A.D. 144, and the same for the cycle of 500 years before.
55. First Two Years of Cycle Beginning 856 B.C.
1st Day of
Lunar Month
Apr. 24, 856
July 22
Sep. 19
Nov. 17
Jan. 15, 855
Mar. 15
May 13
July 11
Sep. 8
Nov. 6
Jan. 4, 854
Mar. 4
56. First Two Years of C
Dec. 20, 357
Mar. 19, 356
May 17
July 15
Sep. 12
Nov. 10
Jan. 8, 355
Mar. 8
May 6
July 4
Sep. 1
Oct. 30
Conjunction
Apr. 25, 3:29 P.M.
July 23, 9:03 A.M.
Sep. 20, 6:52 A.M.
Nov. 18, 4:05 A.M.
Jan. 16, 2:49 A.M.
Mar. 16, 3:09 A.M.
May 14, 7:05 A.M.
July 12, 2:03 P.M.
Sep. 9, 6:19 P.M.
Nov. 7, 7:30 P.M.
Jan. 5, 5:47 P.M.
Mar. 5, 1:15 P.M.
Cycle Beginning 357 B.C.
Dec. 21, 1:27 A.M.
Mar. 19, 1:34 P.M.
May 17, 3:55 A.M.
July 14, 7:32 P.M.
Sep. 11, 8:36 P.M.
Nov. 10, 8:16 A.M.
Jan. 8, 10:27 P.M.
Mar. 9, 3:36 A.M.
May 6, 8:59 P.M.
July 4, 10:47 A.M.
Sep. 1, 4:51 A.M.
Oct. 30, 8:46 A.M.
16
P.M.,
A.M.
A.M.
A.M.
P.M.
P.M.
P.M.
P.M.
P.M.
P.M.
P.M.
P.M.
Morning of
Invisibility
Aug. 16
Nov. 12
Jan. 10
Mar. 11
May 9
July 7
Sep. 4
Nov. 2
Dec. 31
Feb. 28
Apr. 28
June 26
Evening of
Visibility
Aug. 17
Nov. 14
Jan. 12
Mar. 12
May 10
July 8
Sep. 6
Nov. 4
Jan. 1
Mar. 1
Apr. 29
June 28
Morning of
Invisibility
Apr. 25
July 23
Sep. 20
Nov. 18
Jan. 16
Mar. 16
May 14
July 12
Sep. 9
Nov. 7
Jan. 5
Mar. 5
Dec. 20
Mar. 19
May 17
July 14
Sep. 11
Nov. 10
Jan. 8
Mar. 8
May 6
July 4
Sep. 1
Oct. 30
Evening of
Visibility
Apr. 26
July 24
Sep. 21
Nov. 19
Jan. 17
Mar. 17
May 15
July 14
Sep. 11
Nov. 9
Jan. 6
Mar. 6
Dec. 22
Mar. 20
May 18
July 15
Sep. 13
Nov. 12
Jan. 10
Mar. 10
May 7
July 5
Sep. 2
Nov. 1
oi.uchicago.edu
57. In the calculated dates for the cycle beginning 856 B.C., the month starts in every case exactly one day
before the morning of invisibility and two or three days before the evening of crescent visibility. In the calculated
dates for the cycle beginning 357 B.C., ten out of twelve months start on the morning of invisibility, and only one
on the evening of visibility.
58. As a further aid in drawing conclusions from these tables, let us also examine the cycle of 207 B.C., a
year well into the Ptolemaic period and one which begins the seventh of the twenty cycles separating 357 B.C.
and A.D. 144.
59. First Two Years of Cycle Beginning 207 B.C.
1st Day of Morning of Evening of
Lunar Month Conjunction Invisibility Visibility
Nov. 13, 207 Nov. 13, 2:52 P.M. Nov. 13 Nov. 14
Feb. 10, 206 Feb. 10, 8:59 A.M. Feb. 10 Feb. 11
Apr. 10 Apr. 10, 2:00 A.M. Apr. 10 Apr. 11
June 8 June 7, 4:32 P.M. June 7 June 8
Aug. 6 Aug. 5, 1:15 P.M. Aug. 5 Aug. 7
Oct. 4 Oct. 3, 8:29 P.M. Oct. 3 Oct. 5
Dec. 2 Dec. 2, 10:48 A.M, Dec. 2 Dec. 3
Jan. 30, 205 Jan. 30, 8:46 P.M. Jan. 30 Jan. 31
Mar. 29 Mar. 29, 6:33 P.M. Mar. 29 Mar. 30
May 27 May 27, 8:44 A.M. May 27 May 28
July 25 July 25, 12:19 A.M. July 24 July 26
Sep. 22 Sep. 21, 11:13 P.M. Sep. 21 Sep. 23
60. In this last cycle, seven months out of twelve begin with the morning of invisibility, one begins with cres-
cent visibility and the other four are in between. It is clear that, the closer in time the cycle approaches A.D.
144, the less frequently will the months start with invisibility and the more frequently with visibility.
$61. Now it seems incontrovertible that the cyclic calendar must be a regularization of a lunar calendar de-
pending originally on observation. It must, therefore, at its introduction, have been based on the concrete phe-
nomenon either of crescent invisibility or of crescent visibility as the starting point of the month. What we have
now to determine, if possible, is when, or approximately when, the cyclic calendar was introduced. Nothing fits
in 856 B.C., and plainly it could not have been then. If it was around 357 B.C. (plus or minus about fifty years)
then its underlying basis was invisibility. If it was around A.D. 144, then its basis was visibility.
62. We already know, to be sure, that it was in use in A.D. 19, because that cycle is mentioned in the papyrus.
Moreover, the editors state, ". . . aber es kann kaum einem Zweifel unterliegen, dass die in ihm enthaltenen
Rechenverfahren wesentlich alteren Ursprungs sind, da sie keine Spur einer Bericksichtigung der gleichzeitigen
hellenistischen Astronomie zeigen."55
63. A slight bit of evidence in this direction is to be found in the title of one of the books in the Edfu library.
This (from the time of Ptolemy VIII Euergetes II, 145-116 B.C.) is named (Edfou. III, 351): "The
knowledge of the movement (nmtt = cycle ?) of the two lights (sun and moon)."
' 5 6
64. The best possible demonstration, however, that the cyclic calendar was in use before A.D. 19 is to be
found in the double dates (events dated to both the lunar and civil calendars) of Ptolemaic times. These we shall
now examine.
65. DOUBLE DATES INVOLVING LUNAR DAYS. I shall list the dates in chronological order, but with the
latest first and the earliest last, so that we may begin in the period when we know the cyclic calendar was in use
and work back into the preceding period.
66. (1) Thebes. Commodus. Year 30, IIII prt 28 (Al. = II Emw 21 civ. - April 23, A.D. 190) - 1st lunar day.
5 7
67. This date is found on an ostracon wherein one priest leases to another p~y.f 8bd n ht-ntr n st tp n !bd
IIII prt ssw 28 r tpy mw 27, "my temple-month in the first phyle from IIII prt 28 to I mw 27." As we shall see
convincingly under lunar date 8, a temple-month was a lunar month, and we may take IIII prt 28 to be pbdntyw.
Since we cannot be certain whether I mw 27 ends or begins a month, we had better not include it in our present
17
oi.uchicago.edu
discussion. In the publication the year is given as 12(?), but the figure 30 is perfectly clear. The 30th year of
Commodus began July 7, 189, and ended July 6, 190. This would be the 21st year of the cycle beginning in A.D.
169. But in that year a lunar month began not on IIII prt 28 but on IIII prt 22. There remains the possibility that
the ostracon date was given not in the civil calendar but in the Alexandrian. In that calendar IIII prt 28 was April
23, which for A.D. 190 is equivalent to II 9mw 21 civ. This date fits exactly into cycle year 21, and I have no hesi-
tation in accepting it as correct.
68. Calculation:
Apr. 23, A.D. 190 Apr. 22, 8:33 A.M. Apr. 22 Apr. 23
69. (2) Coptos. Nero. Year 12, IIII prt 23 (Al. = I 9mw 15 civ. = April 18, A.D. 66) = 6th lunar day.
5 9
70. On the stela the date of IIII prt 23 is termed p3 hrw n mh 6 n p wr , uthe sixth day of the wrg"; and
wro, as will come out clearly in lunar date 8, means the service in the temple, by lunar months, of the various
phyles. Psdntyw must fall on IIII prt 18. For A.D. 66 this should be in cycle year 22, but there is no agreement.
In the civil calendar, however, the date would be I .mw 10; and this fits nicely in the correct cycle year between
IIII prt 11 and II mw 10.
71. Calculation:
Apr. 13, A.D. 66 Apr. 13, 9:48 A.M. Apr. 13 Apr. 14
72. (3) Thebes. Augustus. Year 21, III mw 10 (Al. = III ,mw 14 civ.
=
July 4, 9 B.C.)= 16th lunar day.
6 0
73. In the hieratic the date appears as =n n I o , "III 'mw 10, being the 16th day,
called hbs-tp." The demotic has simply ibd III mw ssw 10 mh hrw 16 hb-tp, "III mw 10, being the 16th day,
hb -tp." MIller argues that hb -tp is a corruption of hb-sd tpy; and that must be considered a possibility.
6 1
Moreover, when feasts are of long duration, it is usual to mark the successive days as mi hrw 5, ml hrw 6, etc.
However, if this is really the 16th day of a jubilee, the feast began on either June 19, 9 B.C. Al., or June 15, 9 B.C.
civ., and so far as I am aware neither date has any significance in the life of Augustus, either in the present year
or in any preceding one, least of all thirty years earlier.
74. The hieratic version of the date seems to me to speak against this interpretation. If it were to be under-
stood as "III mw 10, called the 16th day of the first jubilee," one would certainly expect the genitive n before
1bs-tp. Much more likely is the view that it is some other kind of 16th day, which itself is named hbs-tp; and that
view becomes almost a certainty when it is recognized that II tmw 25 Al. (16 days prior to III mw 10) is equiva-
lent to II Emw 29 civ., which is the exact date required by year 23 of the cyclic calendar for the cycle beginning
in 32 B.C.
75. One may raise the question why the 16th day of a lunar month should be called hbs-tp, rather than by its
, 63
usual name of mspr sn-nw, Borchardt has explained the name, "Feast of covering the head," as a reference
to the day after full moon, when waning, or
" c o v e r in g ,"
begins.
64
However this may be, we need have only slight
reservation in accepting the date.
76. Calculation:
June 19, 9 B.C. June 19, 6:02 A.M. June 19 June 20
77. (4) Armant. Augustus. Year 1, IIII prt 21 (April 17, 29 B.C.)
=
16th lunar day.
65
78. There is no uncertainty about this date being lunar, as it is given as f - , mspr sn-nw. Psdntyw
is therefore IIII prt 6, and this fits exactly in cycle year 3 of the cycle beginning in 32 B.C.
79. Calculation:
Apr. 2, 29 B.C. Apr. 2, 2:59 P.M. Apr. 2 Apr. 3
18
oi.uchicago.edu
80. (5) Memphis. Cleopatra VII Philopator. Year 8, III 9mw 13 (July 13, 46 B.C.) 5th lunar day.
66
81. The lunar day is given correctly as iht hrwt. Psntyw is III mw 9, which fits exactly in cycle year
11 of the cycle beginning 57 B.C.
82. Calculation:
July 9, 46 B.C. July'8, 5:20 A.M. July 8 July 9
83. (6) Edfu. Ptolemy VIII Euergetes II. Year 30, II Smw (Payni) 9 (July 2, 140 B.C.) = 6th day of lunar
Payni.
67
84. The lunar date is expressed j the 6th lunar day of lhb nt." It is clear that bb mt
1
'(month of) the Feast of the Valley," is merely a variant of p-(n) -int "the one of the Valley," the original of the
name Payni. Psdntyw is II mw 4, which fits exactly in cycle year 17 of the cycle beginning 157 B.C.
85. Calculation:
June 27, 140 B.C. June 27, 9:08 P.M. June 27 June 28
86. (7) Edfu. Ptolemy VIII Euergetes II. Year 28, 1111 mw (Mesore) 18 (September 10, 142 B.C.)
III 9mw (Epiphi) 23 lunar.
6 8
'
87. All three texts cited are necessary to work out the double date. They are given in full below in 214-16
and may be referred to there. Psntyw is III 9mw 26, which fits nicely between II mw 26 and 1111 9mw 25 of
cycle year 15 in a cycle beginning 157 B.C.
88. Calculation:
Aug. 19, 142 B.C. Aug. 17, 11:14 P.M. Aug. 18 Aug. 19
89. (8) Gebelein. Ptolemy VIII Euergetes II. Year 26,
IIII t 20 (Jan. 15, 144 B.C.) = 1st lunar day,
I prt 19 (Feb. 13, o " )=
II prt 19 (Mar. 15, "o)t"
III prt 19 (Apr. 14, " ) "
I111 prt 18 (May 13, "o ) =" "
I m w 1 7 (J u n e 1 ,)
= o
I 9 w 1 J l 1 1, "
) = t "
III mw 17 (Aug. 10, " "
69
90. These dates are to be found in a temple account (Cairo Demotic Papyrus 30801) Since Spiegelberg's
II prt ssw 18 n so tp sw 9 1/4
P--itr sR Pn-t'wy.p3 wy sdt 1 [n p** fy 1n p * h n]ht-ntr
n p.2 wr n Thd II prt ssw 19 r gbd III prt ssw 18
n s* 2 -nw sw 9
19
oi.uchicago.edu
11[..........] ;bd III (or JIll) prt tpy prt SW
50 swn p [. . .] n 3bd IIII1It 55w 20 r tpy prt [ssw 18
. .] Nut-Mn s2 Hp-mn sw 1 swn [ps] Cd n p
13
Cwy 'sdt' (n) drt [X s3 y p3]cwy rsdtl sw 5
IHr-s5-3st s3 Itp-Sbk n p? wrs n Thd III prt ssw 19 r
:bd IIII prt ssw 1714 [n s- 3-nw n p fy 1] n n'
no' blw n n3 wbw n p3 b Mn Iin'p iw n p h n
ht-ntr sw 10
15
Sn-Sbk [s N-nut.f n p3' wr]A n 1bd III prt ssw 18 r tpy
smw ssw 16 n s3 4 -nw n l yncp wph16[
ht -ntr
70
17
n ht-ntr [n pa wr n tpy 9mw ssw 1]7 r 3bd II
smw ssw 16 n s- 5-nw sw 11 1/4
- 18
Sn-Sbk s' Nr-ndt.f n p3 wr 1[n gbd II smw ssw 17 r]
Bbd III .mw ssw 16 n s? tp n p fyWrn
m
ln 3w
n n3 wCbw hnC p h n 11t-ntr
Translation
7
. . The expense of the grain:
8
Horsiese, son of Hetpesobk, (of) the r House of Fire',
from Khelsobk, the prophet of Khonsu, for the grain-
ration and the food of the priests and the expense
of the temple for the service from I prt 19 to
II prt 18 of the first phyle, wheat 9 1/4 (artabas).
P3-ltr, son of Patu, (of) the rHouse of Fire', 10 [for the
grain-ration and the expense of] the temple for the
service from II prt 19 to III prt 18 of the second
phyle, wheat 9 (artabas).
.1[.. III (or IIII) prt? (to) I prt,
wheat 50 (artabas). Price of the [. . .] from
1111 3 ut 20 to I prt12[ 18. .. .. .. Nakhtmin, son
of Hapimen, wheat 1 (artaba). Price of [the] fat
of the r House of Fire', from 13 [X, son of Y, (of) the]
rHouse of Fire', wheat 5 (artabas).
Horsiese, son of Hetpesobk, for the service from III
rain-ra~Tons] and the1foodf he ph;riests ofor +th
from I ?mw 1}7 to II mw 16 of the fifth phyle,
wheat 11 1 /4 (artabas).
20
oi.uchicago.edu
Sensobk, son of Nekhutef, for the service [from II
Smw 17 to] III 9mw 16 of the first phyle, for the
grain-ration and the food of the priests and the
expense of the temple.. .
91. The sequence of total days in each wrS, 29, 30, 30, 29, 29, 30, 30, is conclusive proof that we are here
dealing with lunar months. Dating this document to the 26th year of Ptolemy VIII Euergetes II, whose name
appears nowhere in the text, depends upon the following points:
92. First, the mention of a fifth phyle places it definitely later than the decree of Canopus (237 B.C.), since
that decree reintroduced the fifth phyle;72
93. Second, the paleography of the text places it in the later Ptolemaic period;
94. Third, on the verso of the papyrus, in a different hand, there is an account of grain deliveries, and in
column ii, line 7, appears quite clearly the date, "year 41," which limits the verso to Ptolemy VIII;
95. Fourth, since it was the unvarying practice of demotic scribes to write first upon that side of a sheet of
papyrus which had the horizontal fibers uppermost, the recto of our document must be prior to year 41 of Ptolemy
VIII (130-129 B.C.); and
96. Fifth, the 26th year of Ptolemy VIII (145-144 B.C.) is the first possible year in which the eight lunar dates
will fit into the cycle scheme (year 13 of the cycle beginning 157 B.C.). Four of the dates (IIII t 20, II prt 19,
IIII prt 18, and II smw 17) agree perfectly with those in the 13th year of the cycle, and the remaining four fit nicely
in between.
97. It is not impossible, of course, that the dates are those of a complete cycle earlier, but that would mean
that 40 rather than 15 years fell between the writing on the recto and the verso, which seems quite unlikely from
the rather ephemeral nature of the first text.
98. Calculation:
Jan. 15, 144 B.C. Jan. 15, 8:33 P.M. Jan. 15 Jan. 16
Feb. 13, 144 Feb. 14, 7:50 A.M. Feb. 14 Feb. 15
Mar. 15, 144 Mar. 15, 5:13 P.M. Mar. 15 Mar. 16
Apr. 14, 144 Apr. 14, 1:07 A.M. Apr. 13 Apr. 15
May 13, 144 May 13, 8:17 A.M. May 13 May 14
June 11, 144 June 11, 3:39 P.M. June 11 June 12
July 11, 144 July 11, 12:34 A.M. July 10 July 12
Aug. 10, 144 Aug. 9, 12:05 P.M. Aug. 9 Aug. 10
99. (9) Edfu. Ptolemy IV Philopator. Year 10, III rmw 7 (August 17, 212 B.C.) = 6th lunar day.
7 3
100. This date is exactly 25 years later than lunar date 10, and it marks the close of the first great building
period at Edfu. Psdntyw is III imw 2, which fits properly between II mw 2 and IIII mw 1 in cycle year 20 of
the cycle beginning 232 B.C.
101. Calculation:
Aug. 12, 212 B.C. Aug. 11, 4:47 P.M. Aug. 11 Aug. 12
102. (10) Edfu. Ptolemy III Euergetes I. Year 10, III ~mw 7 (August 23, 237 B.C.) - 6th lunar day.
7 4
103. Instead of III Emw both texts give the month by its early name, 'Ip limt.s, I in the first and\\0
in the second, which is a substitute for L f\ , Ipip, the original of the civil month name Epiphi (230). On
this particular date the cord was stretched to lay out the foundations of the present temple at Edfu, an indication
of the importance of gnt as a building day.
7 5
P4dntyw is III mw 2, and since this date is exactly 25 years
earlier than the preceding one, it falls in the same place in the cycle, year 20, but for a cycle beginning 257 B.C.
21
oi.uchicago.edu
104. Calculation:
1st Day of
Lunar Month
Aug. 18, 237 B.C.
Conjunction
Aug. 18, 1:21 A.M.
105. We have seen that every lunar date of which we have record either is exactly that called for by the cyclic
calendar or fits nicely between the given calendar dates, agreeing with either the preceding or the following date.
This is best seen in Table 4, where all the dates are entered and identified. Observe that the five entries in the
III Amw column are in each case the same as in the preceding II Emw column, while the two entries in the I Emw
column are both the same as in the following column. These latter two both begin, by calculation, on the morning
of invisibility, while four of the five in III mw begin on the day of crescent visibility.
THE DOUBLE
I II III
1
20
9
28
18
7
26
15
4
24
13
2
21
10
30
19
8
27
16
6
25
14
3
22
12
TABLE 4
DATES ENTERED IN THE
PRT
IIII
30
19
8
27
17
6
25
14
3
23
12
1
20(8)
9
29
18
7
26
15
5
24
13
2
21
11
I II
29
18
7
26
16
5
24
13
2
22
11
30
19(8) 19(8)
8
28
17
6
25
14
4
23
12
1
20
10
25-YEAR CYCLE
SMW
III 111
28
17
6
( 4 )
25
15
4
23
12
1
21
10
29
19(8) 18(8)
7
27
16
5
24
13
3
22
11
30
19
9
I II
27
16
5
24
14
3
22
11
30
20
9
28
17(8) 17(8)
6
26
15
4(6)
23
12
2
21(1)
10
( 2 )
10
29
( 3 )
18
8
III 1111
26
15
4
23
13
2
21
10
29
19
9(5) 8
27
17(8) 16
5
26
(7 )
25
14
3
22
11
2(9)(10) 1
20
9
28
17
7
106. Of the entire seventeen calculations, nine months start on the morning of old crescent invisibility, five
on the day of new crescent visibility, two on the day in between, and one on the day before the morning of invisi-
bility. This is precisely the sort of irregularity that one would expect from a cycle scheme, rigidly adhered to,
in contrast to a method of strict observation for starting the month. Most instructive are lunar dates 9 and 10.
They fall in the same place in the scheme; they both therefore begin on the same date in the civil calendar, but
the earlier commences on the day of crescent invisibility and the later (by one cycle of 25 years) on the day of
crescent visibility.
107. In my mind there is not the slightest doubt that the cyclic calendar was in use during the period covered
by the lunar dates. As pointed out in 61, we must choose a time for the introduction of the cycle when it would
22
Morning of
Invisibility
Aug. 18
Evening of
Visibility
Aug. 19
Months:
Year 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
:n C'
oi.uchicago.edu
be in good general agreement with its underlying basis. Crescent visibility as a base must be considered ruled
out by the demonstrated use of the cycle long before A.D. 144, when best agreement on that method would be
obtained. That being so, we are logically driven back to the time around 357 B.C. for the inauguration of the
cycle, at least in the form in which we have it,76 with invisibility of the old crescent as its foundation.
108. CONCLUSION. All the data we have considered-the beginning of the day, the names of the days of the
month and their astronomical basis, the lunar calendar of Pap. Carlsberg 9, and the double dates of the late
period-are in complete harmony with one another in demonstrating the proposition made at the beginning of
this chapter. The Egyptian lunar month, therefore, did begin on that morning when the old crescent could no
longer be seen.
23
oi.uchicago.edu
CHAPTER II
THE LATER LUNAR CALENDAR
5109. FOREWORD. In our discussion of the starting point of the Egyptian lunar month in the last chapter we
incidentally became acquainted with the cyclic lunar calendar of Pap. Carlsberg 9. In the present chapter we shall
investigate the various problems which this calendar poses. It is my hope to demonstrate that we can restore its
missing columns with virtual certainty, that it is not a correction or revision of a previous cyclic calendar, and
that it is a schematization not of the original lunar calendar of Egypt but rather of what may more correctly be
termed the later lunar calendar.
110. THE COMPLETION OF THE CYCLIC CALENDAR. Pap. Carlsberg 9 gives only six dates for each
year of the 25-year cycle. The problem is whether the remaining dates (six or seven) were determined by obser-
vation of the old crescent or by rule, and, if by rule, what that might be.
111. The supposition is that the missing dates were not tied up with observation. The obvious advantage of a
cyclic calendar is that it avoids this necessity, and a calendar which would operate half by rule and half by obser-
vation could not be regarded as a great improvement. There is, however, a stronger basis than supposition for
believing that observation was not used. Of the seventeen calculated lunar dates, nine fall in months for which no
date is given in the calendar scheme (cf. Table 3). Of these nine, only three agree with observation. Most strik-
ing are dates 9 and 10, which fall in cycle year 20, III Smw 2. The first of these, in 237 B.C., agrees with obser-
vation, while the second, exactly 25 years later in 212 B.C., is one day later than observation. Furthermore, if
observation figured in half the calendar, it cannot have escaped attention through the centuries that the scheme as
a whole was getting farther and farther away from reality (107). There seems to be no ground on which to doubt,
then, that the remaining dates were determined by rules. Can we formulate them?
112. To assist us we have the nine lunar dates mentioned above and five certain dates in the I it column. This
is owing to the fact that, in four cases, 60 days (in one other, 90) lie between the dates in IIII mw and II 'ht.
These occurrences fall in the cycle years 5, 10, 15, 20, and 25. Consideration of them and the data summarized
in Table 3 suggests the following rules for determining the unstated dates in the cyclic calendar:
113. Rule I. The dates in I ht are the same as those in II t. This accords with the five known dates in this
column and also starts the calendar off on the first day of the year.
114. Rule II. The dates in III Tht are the same as those in IIIIt. This rule eliminates all but one stretch of
three 30-day months, which falls at the end of cycle year 14 and the beginning of cycle year 15.
115. Rule III. The dates in I prt and III prt are the same as those in II prt. This is based upon lunar dates
8 in cycle year 13.
116. Rule IV. The dates in I Amw and III tmw are the same as those in II bmw. This is based upon lunar
dates 2, 5, 7, 8, 9, and 10.
117. Rule V. Epagomenal dates are the same as those in IIII bmw. This is based upon the one certain date
in cycle year 14, where 90 days fall between IIII mw 5 and II ht 30.
118. Rule VI. In months where two dates occur because of the year having thirteen months, the first agrees
with the previously given columnar date and the second with the following given date. This develops out of Rules
II, III, and IV.
119. The complete calendar scheme is now presented in Table 5. The only column about which I feel any
doubt at all is that of III iht. It lacks a confirmatory date from any source, and one might propose, on the analogy
of the latter part of the calendar, that its dates should agree with those in II lht. We should then have, taking
cycle year 2 as an example, a consistent pattern of 20, 20, 20, 19, 18, 18, 18, 17, 16, 16, 16, 15. This is
plausible and attractive and may, indeed, be correct; but it has the objection that it would result in four stretches
of three 30-day months (in cycle years 4,5, 9-10, 19-20, and 24-25) and one of four 30-day months (in cycle
years 14-15). This situation is not impossible astronomically (18), but it is unlikely that it would be introduced
into a cyclic scheme. Certainty for this column, however, must await a definite confirmatory date.
24
oi.uchicago.edu
TABLE 5
THE COMPLETED 25-YEAR CYCLE
BHT
Months: I II III III
Year 1 1 1 1-30 30
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
20 20
9 9
28 28
18 18
7 7
26 26
15 15
4 4
24 24
13 13
2 2
21 21
10 10
30 30
19 19
8 8
27 27
16 16
6 6
25 25
14 14
3 3
22 22
12 12
19
8
27
17
6
25
14
3
23
12
1
20
9
29
18
7
26
15
5
24
13
2
21
11
19
8
27
17
6
25
14
3
23
12
1
20
9
29
18
7
26
15
5
24
13
2
21
11
I
29
18
7
26
16
5
24
13
2
22
11
1-30
19
8
28
17
6
25
14
4
23
12
1
20
10
PRT
II III
29 29
18 18
7 7
26 26
16 16
5 5
24 24
13 13
2 2
22 22
11 11
30 30
19 19
8 8
28 28
17 17
6 6
25 25
14 14
4 4
23 23
12 12
1 1-30
20 20
10 10
III'
28
17
6
25
15
4
23
12.
1
21
10
29
18
7
27
16
5
24
13
3
22
11
30
19
9
I
27
16
5
24
14
3
22
11
1-30
20
9
28
17
6
26
15
4
23
12
2
21
10
29
18
8
BMW
II III III
27 27 26
16 16 15
5 5 4
24 24 23
14 14 13
3 3 2
22 22 21
11 11 10
30 30 29
20 20 19
9 9 8
28 28 27
17 17 16
6 6 5
26 26 25
15 15 14
4 4 3
23 23 22
12 12 11
2 2 1
21 21 20
10 10 9
29 29 28
18 18 17
8 8 7
120. As a spot check on the agreement of this calendar with observation, I have calculated the derived dates
for the first two years of the cycle beginning 357 B.C.
1st Day of
Lunar Month
Nov. 20, 357 B.C.
Jan. 19, 356 B.C.
Feb, 17
Apr. 17
June 16
Aug. 13
Oct. 12
Dec. 9
Feb. 6, 355 B.C.
Apr. 6
June 5
Aug.-2-
Oct. 1
Conjunction
Nov. 21, 8:33 A.M.
Jan. 19, 4:05 P.M.
Feb. 18, 4:05 A.M.
Apr. 17, 9:07 P.M.
June 15, 10:53 A.M.
Aug. 13, 6:38 A. M.
Oct. 11, 1:22 P.M.
Dec. 10, 3:55 A. M.
Feb. 7, 2:44 P. M.
Apr. 7, 1:22 P. M.
June 5, 3:45 A. M.
Aug. 2, 4:31 P.M.
Sep. 30, 5:20 P. M.
121. Of the twelve given dates for these two cycle years, ten agreed with observation (56). Of the thirteen
derived dates, eight agree with observation. Eighteen agreements out of twenty-five give a percentage of correct-
ness of 72, a high figure, considering the moon's variability and the relative simplicity of the calendar. I should
25
EPAG.
4
2
5
3
1
Morning of
Invisibility
Nov. 2 0
Jan. 19
Feb. 17
Apr. 17
June 15
Aug. 13
Oct. 11
Dec. 9
Feb. 7
Apr. 7
June 5
Aug. 2
Sep. 30
oi.uchicago.edu
think it quite likely that the rest of the calendar is up to the accuracy of the first two years.
122. We shall find that this 25-year cyclic calendar is extremely useful for checking lunar data throughout
Egyptian history. By means of it one lunar datum can easily be compared with another to determine whether or
not they belong together.
123. THE RULE GOVERNING INTERCALATION AND THE POSITION OF THE INTERCALARY MONTH.
The next problem we have to discuss in our treatment of the cyclic lunar calendar is what months were intercalary.
We know of course from Pap. Carlsberg 9 the years of the cycle in which there was an intercalated month (50),
but we are not informed where in the year that month falls.
124. Fortunately, two of the double dates discussed in the last chapter (83, 86) name the lunar month as
well as the civil month. They are:
(7) Ptolemy VIII Euergetes II. Year 28, IIII mw (Mesore) 18 = 23d day of lunar Epiphi.
1
(6) Ptolemy VIII Euergetes II. Year 30, II Emw (Payni) 9 = 6th day of lunar Payni.
1
125. Thus we know for certain that in date 7 lunar Epiphi began on the 26th of civil Epiphi and in date 6 lunar
Payni began on the 4th of civil Payni. It may be taken for granted that it was the intention of the cycle scheme to
have as great a synchronism as possible between the lunar and the corresponding civil months. We have already
noted that date 6 belongs to cycle year 17, and we know that this cycle year is a great year, one in which there is
an intercalation. Unless one wishes to upset completely the partial concord between lunar and civil months in
this year, the only logical intercalary month is the one beginning on the third epagomenal day. It is also apparent
that lunar Thoth of year 15 must have begun on civil Thoth 30, since, as we have seen, lunar Epiphi of this cycle
year began on civil Epiphi 26. The intercalary month in cycle year 14, then, must be the one beginning on the fifth
epagomenal day. Furthermore, since the only months beginning on the epagomenal days are in great years, it is
an obvious conclusion that all months beginning then are intercalary, and the underlying principle must be that an
intercalation is required to prevent the first of lunar Thoth from falling before the first of civil Thoth. If this
were not the principle, we should hardly have lunar Thoth of cycle year 15 beginning on the 30th of civil Thoth,
when without the intercalation in the preceding year lunar Thoth would begin on the fifth epagomenal day and would
thus almost completely synchronize with civil Thoth. It seems incontestable that the beginning of the lunar year
must not precede the beginning of the civil year.
126. Adopting this principle, then, we must place the remaining four intercalations (of cycle years 1, 9, 12,
and 23) in civil Mesore (IIII mw) 26, 29, 27, and 28, respectively. All intercalation would then be based upon a
very simple rule, one which could have been easily applied before the 25-year cycle had been evolved, when inter-
calation was still empirical. Whenever the first day of lunar Thoth would fall before the first day of civil Thoth,
the month is intercalary.
127. There are two other possibilities, however, which must be dealt with. It might be suggested that no lunar
month should begin before its corresponding civil month. If that were true, then in cycle years 1, 9, 12, and 23
we should have to make intercalary either the first or a later month. Assuming the former, note the result in
cycle year 9. The intercalary month would begin on civil Thoth 4. Lunar Thoth itself would begin on civil Paophi
4. Not one day of the first eight lunar months would fall in the corresponding civil months in that year, quite con-
trary to the principle of synchronization between the two to the highest degree possible. Intercalation in the last
month of year 9 would result in good synchronization.
128. The suggestion might then be made that intercalation in these cycle years should be made with the first
month in each dated to the 30th day of a civil month. This would avoid beginning any lunar month before its civil
month, yet at least one day of each lunar month would fall in its civil namesake. The objection to this idea is that,
while it would be theoretically possible once the 25-year cycle had been established, it is inconceivable before
that time, when intercalation was empirical and must have been based on a simple and easily applicable rule.
There is no reason to suppose that the cycle scheme was anything other than the systematization of long-established
practice; that being so, the only simple and practical rule is the one which we formulated above. It is true that
this would result at times in lunar months beginning before their corresponding civil months, but this was of no
consequence so long as the beginning of the lunar year did not anticipate the beginning of the civil year.
129. THE LAG BETWEEN LUNAR AND CIVIL MONTHS. The earliest date for the first of a lunar month is
26
oi.uchicago.edu
day 27 of the preceding civil month. This occurs in year 1, with lunar Mesore (IIII mw) beginning on civil
Epiphi (III Smw) 27. On the other hand, a lunar month may begin as late as the last day of its civil month, as in
year 15, when lunar Thoth (I ibt) begins on civil Thoth 30. Graphically this is represented in Figure 12. The
Range of Lunar Month
Civil Month
FIG. 12.-Possible relation between lunar and civil months.
mean first lunar day would be day 14 of the civil month. Figure 13 illustrates the mean relation between the
lunar and the civil month:
Lunar Month
Civil Month
FIG. 13.-The mean relation between lunar and civil months.
130. We shall return to this lag between the lunar and civil months when we discuss the phenomena from which
Gardiner drew his conclusion that Mesore had once been the first month of the year (282-305).2
131. THE FIRST SCHEMATIC CALENDAR. We have still to consider whether the calendar of Pap. Carlsberg
9 is something new or an older scheme corrected and brought up to date. This is a problem of fundamental impor-
tance for Egyptian chronology, since upon its solution depends our ability to use any earlier lunar datum with
precision.
132. The outstanding fact of the cyclic calendar is that it begins on I it 1. We have come to the decision that
in this form it originated in the fourth century B.C. We have also seen that by A.D. 144 it was late by one day
(54). Had it been desired to correct it at that time, it would have been necessary to lower every date in the cycle
by one day. The cycle would then begin not on I it 1 but on I iht 30, with following dates just like those in cycle
year 15.
133. Despite the fact that in A.D. 144 the cyclic calendar was clearly no longer in agreement with lunar phe-
nomena, it was not corrected. This is certain, since lunar date 1 falls in A.D. 190 and still fits exactly into the
scheme (67). Obviously, then, there was no provision in the calendar itself which required that periodically it be
adjusted. Furthermore, if the present cycle were the result of a correction in the fourth century B.C. of an al-
ready existing cycle, instituted, let us say, 500 years earlier, that earlier cycle would have had to begin with
I Dit 2, so that its correction would result in I iht 1. This is exceedingly unlikely. While the first day of the year
is unquestionably an appropriate day on which to start a calendar, the second, third, and fourth days (as would be
the case with still earlier corrected cycles) have nothing to recommend them. These considerations lead inevitably
to the conclusion that the calendar of Pap. Carlsberg 9 was original with the fourth century.
134. The question may still be legitimately asked, however, whether any sort of schematic lunar calendar was
in earlier use. We have already noted (30) that Borchardt was of the belief that the Egyptians early adopted a
conventionalized lunar calendar of alternating 29- and 30-day months, with now and again one of 31 days. As a
result of this belief he allowed himself a latitude of as many as three days in equating lunar dates with the civil
calendar, and frequently he accepted rather poor equations as being "kalendarisch" or "wohl kalendarisch."
3
With
such a latitude and with the known repetitive character of lunar movement, it is too easily possible to obtain con-
firmation of chronological hypotheses based on lunar data.
135. Borchardt based his schematic lunar year on a temple account of the Middle Kingdom which lists alternate
temple-months of phyle-priests. Since we shall go into this document thoroughly in Excursus C, I shall merely
state here that his interpretation of the dates therein gave him the following sequence of months: 30, 29, 31, 29,
30, 29, 30, 29, 30, 29, 30, with the 31-day month ending on I it 20. Borchardt concluded that the year alter-
nated 29- and 30-day months, and an extra day was added to the month ending in the first civil month in order to
correct it, the implication being clear that the lunar year should start off correctly with an observed month's be-
ginning.
27
oi.uchicago.edu
136. If a 31-day month could be proved (in Excursus C we shall see that it is quite unlikely that the present
papyrus does so), then there would be grounds for acceptance of Borchardt's schematic year. But, even so, it may
be categorically stated that a calendar which began with observation and then alternated 29- and 30-day months
(or 30- and 29-day months) would be bound to average 50 per cent agreement with the correct astronomical begin-
nings of the months, and the difference between the date of any schematic month and the correct date could be as
much as two days on only the rarest of occasions. This is easily demonstrated with actual dates. The first two
years of the cycle beginning in 357 B.C. have been calculated in full (56, 120). In Table 6 the astronomically
correct dates are shown in the central column. The left-hand column consists of schematic dates beginning with
a 29-day month, the right-hand column of schematic dates beginning with a 30-day month; thus both possibilities
are covered. It will be noticed that it is necessary to correct a 29-day month to 30 days when beginning the sec-
ond year in the left-hand column. Since both schematic columns agree in twelve out of twenty-five with the true
dates, the average of 50 per cent agreement claimed above is justified. Moreover, not one schematic date is off
by two days. This in itself would be enough to cast doubt on many of Borchardt's equations.
137. G. H. Wheeler, who offered another interpretation, based on Borchardt's readings, of the same papyrus,
avoided the 31-day month and got instead the sequence of 29, 30, 30, 30, 29, 30, 29, 30, 29, 30, 29.5 He too
believed that he was dealing with a schematic calendar but with the difference that it substituted a 30- for a 29-
day month every 32 months; but this postulates a knowledge of the length of the average synodic month which the
Egyptians could not possibly have had. He stated further: "It is unlikely that the date for celebrating the new moon
would be fixed by a separate astronomical observation every month."6 But that is exactly what was done by all
COMPARISON
Alternating
29 and 30
I.
I
II
III
IIII
I
II
III
IIII
I
II
III
IIII
I
II
III
IIII
I
II
III
IIII
I
II
III
IIII
bt 1
" 30
" 30
" 29
, 29
prt 28
" 28
" 27
S 27
rpw 26
26
" 25
" 25
bt 20
" 20
S 19
S 19
prt 18
S 18
17
17
mw 16
" 16
, 15
S 15
TABLE 6
OF SCHEMATIC AND OBSERVATIONAL YEARS
Astronomically
Correct Dates
= I Tht 1
II " 1
III " 1
III " 30
IIII " 30
I prt 29
II " 29
III " 28
= IIII " 27
I mw 27
- II " 26
III " 26
IIII " 26
S I ht 20
S II " 20
III " 20
S IIII " 19
I prt 19
- II " 18
III " 18
S IIII " 17
= I mw 16
S II " 16
S III " 15
S IIII " 15
Alternating
30 and 29
I ?ht 1
= II " 1
II " 30
= III " 30
IIII " 29
= I prt 29
II " 28
= III " 28
= IIII " 27
= I mw 27
= II " 26
= III " 2 6
IIII " 2 5
= I ht 20
II " 19
III " 19
IIII " 18
I prt 18
II " 17
III " 17
IIII " 16
= I mw 16
II " 15
S III " 15
1111 " 14
28
oi.uchicago.edu
the other ancient peoples who used a lunar calendar. There is no certain proof that even in the latest times did
the Babylonians substitute an arbitrary calendar for observation, and they knew considerably more about astron-
omy than the Egyptians did. Moreover, the lunar year was essentially a religious and priestly year, since the
civil year was available for ordinary purposes, and it is quite legitimate to suppose that conservatism would oper-
ate against the early introduction of a schematic calendar.
138. The dating of Thutmose III's Battle of Megiddo may be offered as evidence of such a calendar. The
passage is:f i n 0, " "Year 23, I mw 21,8 the day of the feast of psdntyw
exactly." Now does this mean that the first day of the astronomical lunar month coincided with the first day of
a schematic lunar month, or does it simply mean that the Battle of Megiddo happened to fall exactly on psdntyw
after that day had been determined by observation? The second interpretation is as plausible as the first, if not
more so.
139. The danger of arguing from the sequence of 29, 30, 30, 30, 29, 30, 29, 30, 29, 30, 29 that this is
merely a schematic alternation of 29- and 30-day months, with one 29 corrected to 30, is emphasized by a con-
sideration of the sequence in lunar date 8, which runs 29, 30, 30, 29, 29, 30, 30 (89). Anyone who knew nothing
of the 25-year cycle would almost certainly conclude that this sequence was the result of observation, not of rule.
In the latter case we know schematic months are involved. In the former all we really know is that the sequence
is one that could be produced by observation. We are completely unjustified, therefore, in assuming that, prior
to the introduction of the calendar of Pap. Carlsberg 9, observation as the means of recognizing the month's be-
ginning was certainly abandoned by the Egyptians. Further, we have no right to allow a latitude of two or three
days in equating lunar dates. While errors in observation remain possible, we must nevertheless, as the only
sound working basis in chronological study, give the greatest weight to that one of two or more possibilities which
agrees with astronomical calculation.
140. THE LATER LUNAR CALENDAR. We have decided that the first schematic calendar was introduced in
the fourth century B.C. and that prior to that time there is no evidence that any other method than observation was
used to begin the month. We have discovered the rule that regulated intercalation in the 25-year cycle and have
seen that it is one which could easily have been operative before the cycle was installed. The essential point to
observe is that the lunar calendar was governed by the civil calendar, since, whenever the first day of lunar
Thoth would fall before the first day of civil Thoth, the month was intercalary (126). Since the civil calendar
moved forward through the natural year, the lunar calendar attached to it must likewise have moved with it. More-
over, this lunar calendar cannot have existed before the civil calendar was introduced.
141. Since it is rightly taken for granted that a lunar calendar was in use in Egypt before the civil year was
inaugurated (33), two questions immediately pose themselves. First, what was the nature of the original lunar
calendar and, second, when and why was the later lunar calendar introduced? An answer to the second question
cannot be hazarded until we have investigated the civil calendar, and that, in turn, must await the answer to the
first question, which we shall attempt to formulate in the next chapter.
29
oi.uchicago.edu
CHAPTER III
THE ORIGINAL LUNAR CALENDAR
.142. EARLIER THEORIES. The belief in the existence of a lunar calendar in Egypt before the civil calendar
was until but recently based almost exclusively on analogy with other primitive peoples and on passages, fre-
quently obscure, in classical writings. As a result of his study of the latter, Gatterer in 1786 had come to the con-
clusion that the Egyptians had three years: one, the civil year; another, fixed to the rising of Sothis (Sirius),
which he termed the "astronomical year"; and a third, a lunar year governed by Sothis. According to him, the
Nile began to overflow at the time of the summer solstice or the new moon. This new moon, called Neith, was
the first one after the day on which Sothis rose heliacally, and as such it was the first new moon of the year.
143. Ideler (1825) thought that probably the Egyptians had originally a lunar year, but he rejected Gatterer's
theory of one tied to Sothis. Since he also rejected the fixed Sothic year, only the civil year remained in his favor.
3
144. Lepsius (1849) accepted, besides the civil year, both the fixed Sothic year and a lunar year which he con-
sidered must have come before the others. He saw traces of this early year in the division of the civil year into
twelve months and the use of a crescent as the hieroglyph for month. On the basis of the classical writers, he
would have the lunar year begin around the summer solstice or the heliacal rising of Sothis.
4
No details were
offered on the individual months or on the method of intercalation.5
145. Martin (1864) and Hincks (1865) continued to maintain the existence of a lunar year tied either to the sum-
mer solstice or to the heliacal rising of Sothis. The former regarded such a year as necessary to keep certain
feasts in their proper places as well as to provide the phases of the moon for astrological speculation. The com-
mencement of this year was kept, by the intercalation of a thirteenth lunar month when necessary, at the last new
moon before the heliacal rising of Sothis, or better at the first new moon after the solstice. With this latter
starting point agreed Hincks, who also, incidentally, credited the Egyptians with a wandering lunar year of twelve
months only, as well as the civil year and the fixed Sothic year.
146. Brugsch (1891) apparently followed Lepsius in accepting an original lunar year beginning around the
summer solstice,8 although earlier (1883) he mentioned only one which ran concurrently with the civil year (the
later lunar calendar).9 Besides these calendar years, Brugsch maintained also the fixed Sothic year as well as
a fixed Canopic year.11
147. Eduard Meyer (1904) recognized that the civil year of the Egyptians was an artificial creation, since
neither month nor season nor even year corresponded to any natural period. Before the introduction of this year
he believed that lunar months and a lunar year of some sort were used, but he offered no theory of such a lunar
year.12 The fixed Sothic year, accepted by so many other chronologers, he rejected.13
148. Sethe's (1920) discussion of the lunar year was complicated by his inability to see it as a preliminary
stage to the year of 365 days. While recognizing that lunar months must have been the earliest measure of time,
he concluded that most plausibly the lunar year and the civil year were parallel in development, that the latter
was not the daughter of the former, but rather her sister.14 As the first stage of the civil year he saw a year of
360 days, with twelve months, after the model of a lunar year, of 30 days each.15 In addition to the fully developed
civil year, Sethe reaffirmed the fixed Sothic year, which would provide a necessary sacred calendar for the natural
feasts.
1 6
149. Borchardt (1935) was the last chronologer to discuss all forms of the Egyptian year, both civil and lunar.
According to him, the first year was lunar, and it began with the next lunar month after the heliacal rising of
Sothis. When, because of the yearly shift forward of the lunar calendar by about eleven days, the first month of
the year would fall before the rising of Sothis, the month was, instead, intercalary. The name of this intercalary
month was wp rnpt. Proof of the existence of such a calendar was to be found in the Ebers calendar and in the
names of certain of the lunar months. After this year had been in use for some time, the civil year was inaugu-
rated, with its beginning marked by the heliacal rising of Sothis. Still later, the lunar year concurrent with the
civil year was developed. By the end of the New Kingdom, or possibly later, the original lunar calendar fell
entirely out of use and was superseded by the later lunar calendar, which was probably present in the Middle
30
oi.uchicago.edu
Kingdom and certainly present in the New Kingdom.
1 7
150. THE PROPOSED ORIGINAL CALENDAR. So that the reader may follow the trend of my own argument
through the pages of discussion ahead, I shall present at this point a statement of what I conceive the original
calendar to have been.
151. Whatever it may have been in prehistory, the first Egyptian calendar of record was lunar, and it was
based upon the heliacal rising of the star Sothis. This event was called by the Egyptians wp rnpt, "Opener of the
Year." The twelve months of the normal year were divided into three seasons, ht, prt, and Emw, of four months
each. The individual months were named after the most important feasts which occurred in them. The first
month of the year, the month of the thy-feast, began with the first day of invisibility of the moon before sunrise
after wp rnpt. This first day of the year was called tpy rnpt. The twelfth month of the year was named wp rnpt
after that feast, which always had to fall in it. Because the lunar year was normally but 354 days long, whenever
the first month began within 11 days of wp rnpt, it was intercalary, lest at the end of that year the feast wp rnpt
fall out of its month. This intercalary month which occurred every three, rarely two, years was dedicated to
Thoth, and a feast of this god, Dhwtyt, was celebrated in it. Figure 14 illustrates the assumed regulation of the
calendar.
wp rnpt
(Rising of Sothis)
L Month XII wp rnpt L Month I tL
wp rnpt L Intercalary
wtyt
ty
.ipt hmt L
wp rnpt
L th, h - -
wp rnpt - - L- th
mnht
wp rnpt L Intercalary D hwtyt t
FIG. 14.-The proposed regulation of the lunar calendar.
152. THE PRIMITIVITY OF A LUNISTELLAR CALENDAR. This type of lunar calendar, one whose beginning
was determined by a star, is by no means unique among primitive peoples. The following quotation from Nilsson
well illustrates this point:
Immediately after the discovery of America it was already reported of certain tribes on the Mexican coast that
they began the year at the setting of the Pleiades and divided it into moon-months. In Loango the months are
counted from new moons, but Sirius, the rainy star, offers a means of correcting the reckoning sidereally. With
the first new moon which sees Sirius rising in the east their new cycle of twelve months begins, and this must run
as well as it can until the new year. When the cycle of months and the year do not fit, which happens about every
three years, a thirteenth month must be inserted. This is the evil time, when the wandering spirits are at their
worst. The Caffres have twelve moon-months with the usual descriptive names: on this account uncertainty often
arises as to which month it really is. The confusion is always rectified by the morning rising of the Pleiades, and
the reckoning goes on smoothly for a time, until the months once more get out of place and it becomes necessary
to refer again to the stars in order to correct them. In Bali the Pleiades and Orion are observed for the purpose
of correcting the calendar of moons by intercalation; thus the month kartika is doubled, or the month asada is pro-
longed until the Pleiades appear at sunset. Moreover, certain natural phenomena are observed. In New Zealand,
where all months were described by stars, the year began with the new moon following on the rising of the winter
star puanga (Rigel); the thirteenth month often passed unobserved, i.e. served as an intercalary month. Elsewhere
we are told that the displacement of the moon-months in relation to the year was rectified through the observation
of the rising of the Pleiades and Orion, and that the most accurate way of calculating the beginning of the year was
to observe the first new moon after the morning rising of Rigel.
1 8
153. Most instructive in the above quotation is the reference to Loango, where the lunar calendar is also based
on Sirius. Loango is on the west coast of Africa, just south of the equator, and it is not impossible that its
calendar either is due to Egyptian influence or is a cultural survival out of the ancient Hamitic substratum of
eastern Africa,
2 0
from which Egypt drew so much and to which we have already pointed in comparing the begin-
ning of the month in Egypt with the beginning of the month among the modern Masais and Wadschaggas (25).
154. Why the Egyptians should have chosen Sothis as the starting point for their year will be discussed in the
next section. Here all it is desired to establish is the fact that the first Egyptian calendar need not have been the
product of a highly developed culture. It had common roots with many other primitive calendars and must be
characterized as quite normal and unspectacular.
155. SOTHIS AND THE INUNDATION. How long a period passed before the formulation of the lunar calendar
as suggested can only be guessed at. There must have been millennia when primitive man used only lunar months
to reckon time; but after the period of food-gathering had begun to end for the earliest Egyptians and the period of
31
oi.uchicago.edu
food-producing had started, the division of time into larger units called "seasons" cannot have failed to impress
itself on them, especially as with the increasing desiccation of the North African plateau more and more people
had to retreat for food and water into the valley of the Nile, where the rhythm of the river from low water through
inundation and high water back to low water again could not be disregarded. Life in the valley divided itself natu-
rally into three seasons, the first of which was the period of inundation (mt), from the time when the river first
began to rise until its level had fallen again enough to expose the surface of the land and permit the sowing of
seed. Following would come the second season (prt) that of seeding, tilling, growth, and harvest. Lastly came
the period of low water ( mw), after the harvest and before the next inundation. Each one of these seasons, primi-
tive man would come to discover, lasted about four lunar months, and over the centuries the idea would grow that
months grouped themselves into seasons, and seasons into a larger unity called "year."
156. The season of inundation, and with it the year, would begin, we may suppose, with the lunar month which
started after the river first began to rise, and the year would then run until the next inundation. But this first rise
of the Nile was, and is, quite a variable phenomenon. In a recent 32-year period (1873-1904) the rise began at
Assuan as early as April 15 and as late as June 23, with the normal beginning in late May or early June.
2 1
During
this same period the smallest number of days from one beginning to the next was 336 days and the greatest was
415 days. A lunar year controlled by the rise of the river might thus have as few as eleven or as many as fourteen
lunar months. When this state of affairs became so distasteful to primitive man that he sought some other basis
of control we cannot hope to know, but as we shall see (172) it must have been at the very latest by the 1st dynasty.
157. The brightest star of all is Sirius, the Egyptian Sothis. In 21 was explained its heliacal rising-the first
day, after its long period of invisibility, when it again blazes forth in the eastern horizon just before sunrise. In
7000 B.C. in the latitude of Assuan, Sothis rose heliacally about May 16 Greg. By 3500 B.C., since the star's
rising is delayed a day about every 120 years, it rose about June 12 Greg. One might surmise, then, that it was
some time in the fifth or fourth millennium B.C., when the rising of Sothis and the normal beginning of the inunda-
tion were quite close together in time, that Sothis came to be recognized as the harbinger of the inundation. It
must have been after this recognition that the calendar was formulated. Primitive man, with the lunar month as
his unit of time, would soon come to the realization that, while the interval between successive floods was highly
variable, the interval between successive risings of Sothis was practically constant. Sothis' rising, then, could be
used as a point of departure for a calendar of lunar months with three seasons, a calendar completely agricultural
and based on the Nile and governed by Sothis only because Sothis itself had come to be the herald of the Nile. A
few decades of trial and error would certainly be sufficient to work out the simple rule of intercalation, so that the
event of wp rnpt would be maintained properly in the last month of the year. Once that had been worked out, the
prehistoric and protohistoric Egyptians would be in possession of a quite adequate calendar. Its variability would
be regarded, at that early time, as a minor factor of little inconvenience, and it always remained in harmony with
nature.
158. THE MOON AND SOTHIS IN TEXTS. There are a few textual references to the moon or Sothis, mainly
from the late period, which connect either or both to a form of year. The following selection, arranged chrono-
logically, may be regarded as typical:
159. (1) Pyramid Texts 965. Osiris is addressed.
"It is Sothis, thy daughter, thy beloved, who has made thy year-offerings in this her name of Year
2 3
160. (2) Hymn to Amon-Re in the temple of Hibeh. Darius I.24
"Amon, the ram, who is in his right eye (= the sun), is in his disk in the sky daily and forever. ... Amon, the
ram, who is in his left eye (= the moon), the moon in the night, the ruler of the stars, who divides day and night,
months, and years .... "
32
oi.uchicago.edu
161. (3) Mariette, Denderah I, 19g =-- Brugsch, Thesaurus, p. 100. It is said of Isis-Sothis
"Years are reckoned from her shining-forth."
162. (4) Temple of Khnum at Esna.
2 5
"One rejoices in the moon (iwn-h
c
) at the beginning of the months; he guides all feasts at the beginning of the
seasons."
163. It is clear that these texts are not decisive as to the kind of year involved. In the absence of any other
data, one could argue from texts 1 and 3, as has indeed been done, that the year concerned was the fixed or
Sothic year, one whose first day always fell on the day of the heliacal rising of Sothis. With this year I shall deal
shortly. For the moment it is merely necessary to point out that a lunar year based on the rising of Sothis would
suit the passages quite as well.
164. THE MEANING OF WP RNPT AND ITS EQUIVALENCE TO PRT SPDT. In our consideration of the
data bearing on this problem we shall do well to begin with the latest period chronologically. The decree of
Canopus affords a good point of departure, as it gives no less than three meanings to wp rnpt:. (1) the birthday of
the king (line 3); (2) the first day of the civil year (line 22); (3) the heliacal rising of Sothis (line 18):
i00 0"on the day of the going-forth of Sothis, called wp rnpt in name in
the writings of the House of Life." We may dismiss the first meaning with the briefest discussion. We shall soon
in these pages investigate the close relationship between the two events wp rnpt and mswt R , "the birth of Re"
(237). Brugsch long ago surmised that what was really meant by wp rnpt in line 3 of the Canopus decree was
mswt R' in the sense of the birth of the earthly Re, the king.
2 7
165. As first day of the civil year, wp rnpt was not uncommon from the Middle Kingdom on. The contracts of
Hepzefi clearly support this interpretation,28 but prior to that period there is no evidence at all to justify it.
Furthermore, we shall see that wp rnpt was in existence before the civil year was inaugurated, so that its appli-
cation to the first day of that year can only have been secondary and through analogy with the lunar calendar.
166. That the primary meaning of wp rnpt was the heliacal rising of Sothis is clear from a number of passages
other than the decree of Canopus. I quote a selection of them:
167. (1) Chassinat, Dendera, III, 35. Hathor is addressed as
"Sothis in the sky, sovereign of stars, rising in the sky on wp rnpt."
168. (2) The astronomical ceiling at Dendera (Brugsch, Thesaurus, p. 10). Sothis is depicted as a recumbent
cow in a boat.
"Sothis the great, lady of wp rnpt."
169. (3) The Ramesseum astronomical ceiling (P1. II LD, III, 171-170).
"Thou (Ramses II) risest like Isis-Sothis in the sky on the morning of wp rnpt."
170. (4) The renewal by Thutmose III of Sesostris III's calendar of offerings in the temple of Semneh (LD, III,
55, a, 11. 8-11 = Brugsch, Thesaurus, pp. 399-400 - Urk., IV, 193-96) lists six feasts in what appears to be
chronological order: (a) tp trw, "feast of the beginning-of the season," which must have fallen on I lht 1; (b) tp trw,
which must have fallen on I prt 1; (c) wp rnpt; (d) hb gbf 'Iwntyw, "feast of repelling the Troglodytes," on IIII prt
21; (e) tp trw, which must have fallen on I smw 1; (f) feast of I smw 1. Since the festival of wp rnpt fell between
33
oi.uchicago.edu
I prt 1 and IIII prt 21, it obviously could not have been the first day of the civil year. We have, however; only to
recall that in the seventh year of Sesostris III prt Spdt, "the going-forth of Sothis," took place on IIII prt 1629 to
conclude that wp rnpt here must mean prt Spdt.
3 0
171. (5) In the tomb of Khnumhotep occur two lists of festivals (Beni Hasan, I, 53-54 and P1. XXIV; ibid., p. 61
and P1. XXV; both in Brugsch, Thesaurus, pp. 231-34). The second includes wp rnpt. The first omits wp rnpt but
includes prt Spdt. While the two lists are not virtual duplicates, it seems clear that the-same festival is meant
by both wp rnpt and prt Spdt.
3 1
172. (6) The right half of a tablet of the 1st dynasty (Petrie, Royal Tombs, Vol. II, Pls. V, 1, VIa, 2)32 bears
the figure of a recumbent cow (Isis-Sothis)33 which has the sign for "year" with a stroke between its horns.
(Figure 15). Underneath is a sign which is apparently _t; and the whole can be plausibly read: "Sothis, the
opener of the year; the inundation."
Ii'4
FIG. 15.-Ivory tablet of the first dynasty.
173. Before considering the implications of the above passages, it remains to be pointed out that prt Spdt is
not mentioned, so far as I know, either as a feast or as an astronomical event, before the Middle Kingdom. It does
not occur in the lists of feasts in the Old Kingdom mastabas which we shall shortly discuss, nor have I found it
on any stelas of the first Intermediate Period.34 In the Middle Kingdom it does appear in tomb inscriptions (cf.
Beni Hasan above) and stelas, though not together with wp rnpt.35 When, later, it does appear with wp rnpt, we
can feel sure that here wp rnpt means the first day of the civil year.36
174. It cannot, I believe, be sheer coincidence that the Middle Kingdom was the time when the term prt Spdt
first appeared and when wp rnpt came to mean the first day of the civil year... The explanation is to be found, as I
see it, in the transfer of wp rnpt from its original special application. As "Opener of the Year" it would mean the
heliacal rising of Sothis, assuming a lunar year based on Sothis. Thus, as a specific day, it did control, or open,
the lunar year. When, however, the civil year had been developed, there came a time, the Middle Kingdom by all
the evidence, when the first day of that year, the day which literally "opened" it, also came to be called wp rnpt.
3 7
Here the emphasis was on the day itself, not on any astronomical event which took place on it. Since by then the
civil year had supplanted the lunar year in the life of the people, wp rnpt as the rising of Sothis was of interest
mainly to the priests and the temples, and another term, prt Spdt, "the going-forth of Sothis," purely descriptive
of the event, was adopted to name the feast in the civil calendar.
175. So far we have dealt in the main with generalities. We have established the reasonableness of a primitive
lunar calendar based on the rising of Sothis and also the fact that this rising was termed wp rnpt. In the following
pages I shall present the evidence which has led me to the conclusion that the year which was opened by Sothis'
rising cannot have been the civil year or the fixed Sothic year but must have been the natural lunar year.
176. THE LISTS OF FEASTS IN THE MASTABAS OF THE OLD KINGDOM. Beginning in the 4th dynasty
the mastabas of the Old Kingdom frequently exhibit a .htp-di-nsw formula invoking Anubis (at times Osiris) and re-
questing prt hrw, "invocation offerings," on certain festivals.
3 8
If a number of these lists are examined, it will
be found that the feasts tend to follow a definite order. A convenient source for this investigation is Mariette,
Les mastabas de l'ancien empire, pub....par G. Maspero (Paris, 1889), and a summation o
f
the 4th and 5th
dynasty lists is presented in the following table.
34
oi.uchicago.edu
(1) wp rnpt
(2) Dhwtyt
(3) tpy rnpt
(4) wig
Order of Feasts
(5) hb Skr
(6) hb wr
(7) rkh
(8) prt Mn
Mastaba No.
4th Dynasty
C 3
C .9
C 18
C 21
C 22
5th Dynasty
D 10
D 12
D 16
D 19
D 23
D 24
D 28
D 28
D 39
D 40
D47
D 48
D 59
D 60
D 62
D 67
*The parentheses inclose elements
not be present in the name.
Feasts
(9)
(10)
(11)
(12)
Listed
( bd)* (n) S/d
(tp) gbd
tp smdt
hb nb rc nb or variant
All
All
1-6
All
1 - 6; 10 - 12
1 - 6; 12
1 - 4; 6 - 12
1-5
All
1-9
1 -9; 12
1 - 11
All
All
1-7; 10-12
1 -9; 12
1 -4; 12
1-8
1-6; 8-9
1 - 6; 12
which may or may
177. The exceptions to the foregoing to be noted in Mariette's work are:
Mastaba No. Order of Listed Feasts
4th Dynasty
C 27 10, 11, 1, 2
5th Dynasty
D52 4,2
D 60* 1, 3,4,2, 10, 11,12
D 61 3, 1, 2, 12
D 69 1,4,2,3,5,6,7,10,11
D 38 has been omitted in both tables
because of lacunae.
*See also preceding table.
178. The exceptions to be found are certainly insufficient to weaken the overwhelming evidence of a strict
order to the calendar of feasts in which the dead expected to take part. This order, it is easily demonstrated, can
be nothing other than chronological. Whatever may be the exact meaning of wp rnpt and tpy rnpt, they clearly be-
long at the head of a list. If we now check feasts 4 to 8 against the later temple calendars
3 9
(Medinet Habu and
the Greco-Roman temples), we find the followin dates in the civil year on which they were celebrated:
35
oi.uchicago.edu
(4) wag I bt 18
(5) hb Skr IIII ;ht 26
(6) hb wr II prt 440
(7) rkh (wr) II prt 9 (Edfu)
III prt 1 (Illahun)
4 1
(8) prt Mn I smw 11 (Med. Habu)
42
Nos. 9 - 11 were monthly feasts, celebrated at least twelve times a year.
179. There seems full justification for considering the Old Kingdom lists to be arranged chronologically. More-
over, there are reasons which will come out in the following pages for thinking that all the feasts, with, of course,
the exception of wp rnpt, were lunar at that time.
4 3
The wag-feast can be the movable lunar feast for which there
is evidence from the Middle Kingdom (182); rkh as the name of a lunar month cannot be other than lunar (230);
and the lunar character of prt Mn is brought out in the Medinet Habu calendar and later (204,240). The monthly
feasts of bd and 'mdt, and so probably of s d were lunar. Furthermore, there is no other plausible explanation
for the sequence wp rnpt, Dhwtyt, and tpy rnpt than the assumption that the latter two also were lunar.
180. The proposed original lunar calendar fits the chronological order perfectly, and I know of no other explana-
tion. Wp rnpt was the rising of Sothis, the event which opened the new year but which, in itself, did not form part
of it. Tpy rnpt was the first day of the new year, the first day of the month thy in which fell the feast of wag; and
the remaining feasts followed in chronological order. As for the feast of Dhwtyt, between wp rnpt and tpy rnpt,
this can be nothing other than the feast of the intercalary month which would occur at three- (at times two-) year
intervals. As a special month it was fittingly dedicated to Thoth, the moon-god.
181. The one objection that might be brought against the proposed calendar has to do with this feast of Thoth.
In the Medinet Habu and later calendars, it is listed on I :bt 19, following the feast of wg on I ht 18 and preced-
ing that of thy on I bt 20. How is this to be explained? We shall investigate later in more detail how certain
feasts of the lunar calendar were fixed in terms of the civil calendar. We shall see in the next section how the
feast of wg was fixed to the 18th day of the first month of the civil year, with the result that in the Middle King-
dom there were two such feasts, one fixed and the other movable. This same treatment must have been accorded
many other feasts. But the feast of Thoth fell in an intercalary lunar month which had no counterpart in the twelve
months of the civil year. What would be more natural than to assign it a place in the first month? In late times
we know that this feast had so superseded th in popularity that it gave its name to the month in which it fell.
182. MIDDLE KINGDOM DATES OF THE W3G-FEAST. The Illahun papyri reveal clearly that, in addition
to the wg-feast celebrated on the 18th day of the first civil month, there was a movable wg-feast, which in the
examples known to Borchardt fell on various days in the tenth month of the civil year.
4 4
The clearest and most
complete example dates the feast to II *mw 17 in year 18 of a king who can be none other than Sesostris III or
Amenemhet
III.45
183. The assumption is, of course, that this movable feast fell on a certain day in the first month of the lunar
year. The possibilities for the two kings may be tabulated as follows:
4 6
Sesostris III Amenemhet III
Year 18 = 1861 Year 18 - 1825
wp rnpt
=
prt Spdt IIII prt 19 IIII prt 28
1st lunar month after begins I Emw 4 I Emw 5
2d .. " . . II Emw 4 II Emw 5
wig-feast II Emw 17 II Emw 17
According to Borchardt, both results would be satisfactory; he would call the first lunar month in each case an
intercalary month.47 But, according to his own definition (149) "die Anfbinge aller Schaltmonate liegen also hier
auf oder vor dem Hundssternfrihaufgang." Apparently he did not notice the contradiction.
184. Our proposed lunar calendar, however, would require an intercalation in the case of Amenemhet III only.
Since the next lunar month began only seven days after wp rnpt, without an intercalation the twelfth month of the
succeeding year, wp rnpt, would fail to include its eponymous feast. The month beginning II mw 5 would properly
be the first month of the year, thy, and the wig-feast would fall within it. No such case can be made out for the
36
oi.uchicago.edu
18th year of Sesostris III; there is no plausible explanation of II smw 17 as the date of a wg-feast in 1861 B.C.
185. One further check substantiates placing the date under Amenemhet III. The t y-feast, which took place
48 49
in the civil calendar two days after the wg-feast, was probably originally a full-moon feast. It is a natural
assumption that in the lunar calendar also it fell two days after the wg-feast. In the case of Amenemhet III this
would be II smw 19, the 15th day of the month and the full-moon day. In the case of Sesostris III, II smw 19 would
be the 16th day of the lunar month.
186. THE TEMPLE YEAR AT ILLAHUN. We have already had occasion to mention the temple account of
the Middle Kingdom which lists alternating months of phyle-priests (135). The six months enumerated began
with II smw 26, continued with IIII mw 25, II fht 20, IIII gtrl19', II prt 18, and ended with IIII prt 17. Since
a twelve-month period is covered, the suggestion is strong that some sort of lunar year is involved. The six alter-
nating months must then be either months 1, 3, 5, 7, 9, 11 or months 2, 4, 6, 8, 10, 12 of that year. The result of
our calculations in Excursus C will be to fix the account to 1813-1812 B.C. At that time wp rnpt
=
prt Spdt took
place on I smw 1. This day is included within the last month listed, IIII prt 17 to I smw 16; and this month
would, according to the proposed original lunar calendar, be that of wp rnpt, the twelfth month of the year. The
year began, then, one month before II smw 26, on I smw 27 or 26, correctly the first month of the lunar year
following after wp rnpt on I smw 1. The original lunar calendar would seem, therefore, to have been the basis
for keeping account of the temple service in the Middle Kingdom, at least at Illahun; for that calendar does afford
an explanation of why an account covering twelve months should begin on II smw 26.
187. Since I have been following a chronological order in the presentation of the data bearing on the nature of
the original lunar calendar, a most controversial piece of evidence, the Ebers calendar, must now be dealt with.
Under this topic a discussion of the fixed year is mandatory and cannot longer be postponed.
188. THE EBERS CALENDAR AND THE FIXED OR SOTHIC YEAR. Since a photographic facsimile of the
calendar is readily available in Borchardt, Mittel (Blatt 1 opposite p. 20), I shall present only a hieroglyphic tran-
scription (Figure 16), which should be adequate for discussion purposes.
189. This famous calendar, at the beginning of the verso of the medical Papyrus Ebers, has, since its initial
publication in 1870 by Brugsch,
5 0
been the subject of innumerable chronological disquisitions. Prior to Borchardt's
theory as set forth in Mittel, pages 19-29, it had come to be accepted that the first two lines of the calendar were
to be translated: "Year 9 under the majesty of the King of Upper and Lower Egypt, Dsr-ki-r (Amenhotep I), may
he live forever. New Year's day, III 'mw 9, the going-forth of Sothis. "
5 1
Borchardt, however, came to quite a
different interpretation. His conclusions were: (1) the names which appear in the first column are those of lunar
months; (2) the first name, wp rnpt, is that of the intercalary month, and following it come the month names in the
same order as we know them in Persian times (although some of the names are changed); (3) the repeated sign in
the third column, which follows the designations of months of the civil year, is an abbreviation for psdntyw and is
consequently to be read "new month day" and not "day 9"; (4) the purport of the first two lines of the calendar is
that in the ninth year of Amenhotep I the intercalary lunar month wp rnpt began on a certain "new month day" which
fell in the third month of the third season of the civil calendar, and that on this same day happened the rising of
Sothis.
190. A criticism of Borchardt's interpretation has been offered by W. F. Edgerton.
5 2
He limits himself to a
consideration of the hieratic group which Borchardt wished to read as an abbreviation for psdntyw, objecting that
no exact parallels for such a writing were cited
5 3
and that, while it is the normal hieratic group for 9, it is also
(contra Borchardt) the exact form which was used for "9th day of the month" at that time. His objection is co-
gent, and Borchardt's interpretation of "day 9" as psdntyw must be considered highly unlikely. What, however, of
the rest of his theory? May it still be true that the first column lists lunar months?
191. As a starting point, it is clear that the calendar deals with two kinds of years, the second of which must
be the civil year, since it dates the rising of Sothis to III Amw 9 and it is inconceivable that Amenhotep I reigned
at such a time that this event could be dated so in any other calendar ever devised by chronologers. Leaving to
one side for the moment the identification of the other year, in the first column, we must consider the purpose of
the calendar, which is to present a table of correspondence between the months of the first column and the civil
year.55 The necessity for this table is to be found in the contents of the papyrus, since certain recipes therein
37
oi.uchicago.edu
are restricted in use to certain periods of the year.56 An obvious conclusion, then, is that the papyrus year is that
of the first colum, since if the civil year were intended no correspondence table would be necessary.
A --- C,
FIG. 16.-The Ebers calendar.
that the civil year was not inaugurated until after ca. 2937 B.C., a date later than that to which I tentatively assign
the reigns of Atothis and Usaphais.
5 9
Since it is highly unlikely that a fixed Sothic year could have existed before
the civil calendar, it is possible to conclude that when the recipes of the medical papyrus were being first com-
38
oi.uchicago.edu
piled the only calendar known was the original lunar calendar. This, however, may be regarded as an argument
of doubtful force, and I could not press it without an extensive investigation into early dynastic chronology which
I am not prepared to undertake in this study. The arguments against the existence of a Sothic year must come,
then, from other sources.
194. Long ago Brugsch, who believed in a Sothic year, nevertheless felt compelled to admit that in all the
inscriptions known to him, the decree of Canopus excepted, there was neither a word nor a sign which with cer-
tainty could be taken to name a 6th epagomenal day.
6 0
Nor has any such thing come to light since Brugsch's day.
Moreover, it is the opinion of H. E. Winlock, in his recent study of calendar development in Egypt, that
. . the ancient Egyptians, from the Old Kingdom to the Roman Period, have not left a single trace of such a
fixed calendar. Out of the thousands which have survived from dynastic Egypt, not one document gives equivalent
dates in the known "wandering" year and the hypothetical "fixed" year. Furthermore, by the time that relations
with the outside world were such as to result in unprejudiced foreign evidence on the customs of Egypt, we find
the Egyptian both ignorant of, and unreceptive to, the idea.
62
195. Proponents of the fixed year would argue that the Ebers calendar invalidates the first part of this quota-
tion, and unfortunately Winlock does not discuss the point; but it is true that the evidence from the Greek period
speaks against the fixed year. The decree of Canopus, which was an attempt to arrest the forward movement of
the civil year as of 237 B.C. by the introduction of a 6th epagomenal day every fourth year, militates against the
existence of a fixed year before that time; and this argument is reinforced by the fact that the decree was of abso-
lutely no effect. The civil year continued on uninterruptedly, and the first undisputed fixed year in Egypt was the
Alexandrian, which Augustus began (23).63
196. Despite the lack of evidence from the late period and the general presumption against a fixed year, its
most recent protagonists, Sethe and Weill, have attempted to maintain it as a necessity for seasonal feasts. It
is, of course, clear that a lunar year based on Sothis provides a calendar for the natural or agricultural year
which is just as adequate as a fixed calendar based on Sothis. I do not believe that Sethe and Weill would have
supported the latter, had they been aware of the existence of the former, especially since the arguments they
advanced in favor of the fixed year, which we shall now consider in detail, are quite indecisive.
197. (1) The offerings to the Nile at Silsile, to be made on III 9mw 15 and I 'ht 15, apparently at the period of
low water and the beginning of the inundation, respectively, were founded in the first year of Ramses II, were con-
firmed by his son some 66 years later, and were again confirmed by Ramses III after another 40 years.
198. The first year of Ramses II (ca. 1300 B.C.) was only a few years away from the time when the rising of
Sothis had fallen on the first day of the civil year (ca. 1313 B.C.), so that the civil calendar and the natural year
were in harmony. A period of 106 years would cause a shift in the civil year of only 26 days, one which hardly
required correction in view of the wide variability (as much as 80 days)65 of the beginning of the inundation.
199. (2) The festival of the harvest goddess Renutet was celebrated on I smw 1. If this date were in the civil
calendar, the harvest feast would move throughout the year, instead of remaining in the harvest season.66
200. Besides the feast of Renutet named in the Esna calendar on I smw 1, there was another feast to the same
goddess on I smw 25.
6 7
The likely conclusion is that there was a movable feast as well as a fixed feast, just as
there was a movable and a fixed wig-feast in the Middle Kingdom. The feast fixed to I smw 1 in the civil calendar
would revolve around the natural year; but the movable feast, movable only in terms of a date in the civil calendar,
would remain in place as far as the harvest was concerned if it was determined by the original lunar calendar. We
shall see later (S242-49) that the date of the Esna calendar is such as to lend support to such a conclusion.
201. (3) The Edfu calendar lists on I ~mw 1 a feast which was apparently a harvest festival, as in part of the
ceremony Harsamtawi trod under foot grains of barley.
6 8
202. Sethe failed to mention that this festival, which was the voyage of Horus to Dendera, began on the first
day of the lunar month I Emw. The journey began on psdntyw, and on the subsequent day of full moon there was a
great feast in the entire land. In the original lunar calendar I Emw would fall in harvest time.
203. (4) Part of the ceremonies of the feast of Min required that the king cut a sheaf of grain. The Medinet
Habu calendar places the feast in I Amw 11.69
204. Again Sethe failed to note that it is a lunar date which is involved. The heading over the reliefs on the
feast (MH IV 197 : 1) gives no day date and merely states:
39
oi.uchicago.edu
"I Emw, in it occurs the feast of Min." The calendar, however (MIH, III, 167 1430), reads in full:
"I smw 11, the day of the procession of Min to the terrace, the psdntyw-feast being in the morning." Moreover,
it is well known that Ramses III's calendar is a copy of Ramses II's, and we have noted in 198 that in the reign
of Ramses II the civil calendar and the natural year were in harmony. That being the case, any harvest feast in
the ninth month of the civil year would be quite proper, without any lunar calendar being involved; but, as we shall
shortly see (238-41), the feast of Min was in truth regulated by the original lunar calendar.
205. (5) The Medinet Habu calendar is a striking example of the fixed year because the rising of Sothis is
dated in it to I ;t 1.70
206. This is an error of Sethe's which he took over from Brugsch. The correct entry (MH, III, 152 : 629) is
"I it, when Sothis goes forth on her day." This could be any day in I St from 1 to 30, from ca. 1313 B.C. to
ca. 1193 B.C. in the civil calendar. That range includes the whole reign of Ramses II, whose calendar was the
prototype of Ramses III's, and part of the reign of Ramses III himself.
207. (6) The Karnak water clock of Amenhotep III is based on a year beginning in the summer, which must be
the fixed Sothic year.71
208. The calibrations on the interior of the water clock are shown in Figure 17.72 The theory of the clock is
quite simple. It was filled to the brim with water at sunset. When the water, flowing out slowly through an out-
let in the bottom of the clock, had dropped in level to the first mark of the appropriate month-scale, the second
hour of the night began. The shortest scale is that of II 9mw and the longest IIII 3t. From this one may conclude
that I 3ht 1 was near the autumn equinox, about October 5 in the time of Amenhotep III. It is perfectly clear, then,
that the same day cannot have marked the rising of Sothis, since that event would take place about July 19. Con-
sequently, the water clock is no example at all of a fixed Sothic year.73
3ht prt
I II III IIII I II III IIII I II III IIII
.* . . .* . .0
" " " " " " " " " "
" " " " " " " " " "
" " " " " S
209. (7) Plutarch, using the Alexandrian year, put the feasts of Osiris in the month of Athyr; and a correct
equation with the month of Choiak, the proper month for the feasts, can only be made by having that month in a
40
oi.uchicago.edu
Sothic year. Then Athyr 17-19 Al. would be equivalent to Choiak 28-30 Sothic.
7 4
210. The festivals of Osiris in the month of Choiak began on the 12th and ended on the 30th with the erection
of the djed-pillar. According to Plutarch, Athyr 17 marked the death of Osiris, and this certainly took place
before Choiak 28. Moreover, Plutarch speaks of the ceremonies as extending over four days, but begins them on
the 17th and concludes them on the 19th.
7 6
It is rather temerarious, therefore, to insist on the equation Athyr 19
Al. = Choiak 30 Sothic, when the correctness of the first element is in doubt and also when the four days of Plutarch
may quite legitimately fall elsewhere in the period Choiak 12-30 than at the very end. Loret, at the close of his
study of the Dendera texts on the Osirian feasts, concluded that Plutarch had merely mistaken the month of
Athyr for that of Choiak. While this must remain a possibility, what seems the more likely explanation is that
the civil year was nearly the same as the "Sothic" year in the latter part of Plutarch's life (ca. A.D. 46 - ca. 120)
and that his dates are simply those of the civil year rendered in terms of the Alexandrian calendar. Assuming
that his four-day festival actually began on the 16th of Athyr, Athyr 16 Al. would be equivalent to Choiak 12 civ. in
A.D. 76. Assuming that his festival actually ended on the 20th of Athyr, Athyr 20 Al. would be the same day as
Choiak 30 civ. in A.D. 132. It is probable that his dates reflect a year which fell somewhere between A.D. 76 and
the end of his life, and there is no need to postulate a Sothic year to account for them.
211. To recapitulate, every argument that has been brought forward by Sethe and Weill to justify a fixed year
is either based on incorrect information or is susceptible of another explanation. It is impossible to maintain the
Ebers calendar as the sole evidence for the fixed year, since it too is susceptible of another explanation, which is
that it is a table of concordance between the original lunar calendar and the civil year. There is plenty of clear
evidence for the original lunar calendar and none at all for a fixed year; and since the Ebers calendar can have
been made up with the former just as well as the latter, one would be foolhardy indeed to consider it proof of a
fixed year. ,
212. There are still difficulties, however, to be met in properly interpreting the first column. We have seen
that Borchardt considered the month wp rnpt to be the intercalary month (though that would mean no listing of a
twelfth month), as it headed the list and relegated to second place the month thy, which ought to be first. Further-
more, he dismissed as a "Fehlschluss" the conclusion of Gardiner,
7 8
following Brugsch,
7 9
that wp rnpt in an Edfu
text is a variant name of the last month of the year, usually called Mesore (mswt RC).
8 0
213. In this last he was clearly wrong. The astronomical ceiling in the tomb of Senmut (220-23 and Pl. I)
gives the name of the last month of the year as wp rnpt, as do also the geographical papyrus of Tanis (Figure 18)81
and the fragment of a water clock of Necho found recently at Tanis. Moreover, a comparison of all the pertinent
Edfu material leads to the same result.
FIG. 18.-The fragment of the geographical papyrus of Tanis
which names the last month of the year wp rnpt.
41
oi.uchicago.edu
214. (a) Edfou, VII, 7.
0n i n o N N. nnnn
"Year 28, day 18 of the 4th month of smw (Mesore) under Ptolemy VIII Euergetes II . . ., making 95 years
since the foundation to the feast of entering it."
215. (b) Edfou, IV, 8-9.
/ n .N.N.... I nnn
"Year [28], day 18 of wp rnpt . . . Ptolemy VIII Euergetes II ... , making 95 years from the stretching of the
cord to the feast of entering it."
216. (c) Edfou, IV, 2.83
tA" _.11 1 II - 0I ,-, - ---
"Day 18 of the 4th month of smw (Mesore), being day 23 of the 3d (lunar) month of smw (Epiphi), this beautiful
day of the feast of entering it."
217. These three texts unquestionably refer to one and the same event, the completion and dedication of the
main part of the temple of Edfu, and the equation between IIII mw (Mesore) and wp rnpt is undeniable. In view of
(c) it is impossible to consider wp rnpt a lunar month here, as Borchardt wished to do. Moreover, since wp rnpt
is a variant for Mesore, it cannot be the name of an intercalary month. Why then does it head the Ebers calendar?
218. The explanation is not complicated. The event which regulated the original lunar calendar was the rising
of Sothis, called wp rnpt. The date of this event would, then, correctly go at the head of a calendar governed by it.
But this event also gave its name to the last month of the year. In the first column of the Ebers calendar, there-
fore, the last month of the year appears at the head of the months merely because its eponymous feast determined
the following year.
8 4
The correct interpretation of the second line of the calendar seems to me to be that the date
III 9mw 9 is common both to the going-forth of Sothis and to the beginning of the lunar month wp rnpt. From
this date as a starting-point was projected a schematic lunar calendar of full months of 30 days.
8 6
The failure of
the scribe to reckon the epagomenal days is accordingly not deserving of blame, as it would be were we concerned
with a fixed year. The schematic lunar calendar was regarded as exactly that; it was to be merely a guide to the
proper identification of the lunar month which included the 9th day of any civil month in its first few days.
219. On the basis of the original lunar calendar it is thus possible to offer a logical and consistent explanation
of the Ebers calendar, with no necessary assumption of scribal error or carelessness.
220. THE ASTRONOMICAL CEILING IN THE TOMB OF SENMUT. In 1926/27 the expedition of the Metro-
politan Museum of Art uncovered an unfinished tomb of the well known 18th dynasty figure, Senmut.87 By far the
most interesting feature of the one decorated room is its ceiling (P1. I). In the southern half (above in P1. I) the
decans are listed, while in the northern half are depicted the northern constellations in a central field and beside
them twelve circles bearing the same names88 that appear in the first column of the Ebers calendar. The upper
row of circles runs from right to left and the lower from left to right. Each circle is divided into twenty-four seg-
ments. Under the circles are two opposing rows of deities bearing disks on their heads.
221. It is my conviction that the circles represent the eponymous monthly feasts of the original lunar calendar,
with the twenty-four segments each an hour of the feast day. The thy-feast is here correctly in the first place and
wp rnpt in the last. A division of the year into the three seasons is indicated by the fact that months 1 - 4 are above
and to the right of the central field, months 5 - 8 are to the left, and months 9 - 12 to the lower right. If the seasons
were not intended, symmetry would dictate that six circles be on each side.
222. It cannot be argued that the circles represent the civil months and the civil year. Not only does the Ebers
calendar speak against that, but the clearest possible evidence that we are here concerned with a lunar calendar
is the fact that the deities below the circles are deities of the days of the lunar month. A comparison of the gods
42
oi.uchicago.edu
named here with those of the lunar months can easily be made by referring to Plate V or to Brugsch, Thesaurus,
pages 46-48, where are summarized the Greco-Roman lists. Behind Isis on the right appear in order the gods
of days 4 - 10; an eighth deity, hkw, is not found in the late period. On the left ir m cw* is the god of day 15, tknw
of day 13, *d hrw of day 16, nhs of day 30; the remaining three are not in the late lists. Considering that the
ceiling is more than a thousand years earlier than the late lists, the fact that four out of fifteen are unidentified
is not enough to outweigh the strong presumption that they also are earlier lunar day deities who have been sup-
planted.
223. Further support to the theory that the Senmut ceiling has to do with a lunar calendar is its analogy to the
Ramesseum astronomical ceiling, to which we now turn. This ceiling, as we shall see, has an arrangement which
is a strong indication of such a calendar, above and beyond what Senmut's has to offer.
224. THE ASTRONOMICAL CEILING OF THE RAMESSEUM. Superficially this well-known ceiling with its
balanced arrangement (Pls. II-III) bears slight resemblance to that of Senmut, but closer scrutiny shows that all
the elements present in the one are also in the other, though partly in altered form. The Ramesseum ceiling is
divided into three horizontal panels, the upper one of which corresponds to that half of the Senmut ceiling which
lists the decans. The central band comprises the northern constellations flanked by the two opposing rows of
lunar day deities which appear on the Senmut ceiling under the circles, with a few additions not found in the late
lists. In the lowest panel, where we might expect the circles, there is the greatest change. Ramses II is shown
making offering to various deities who represent the months. Eight out of twelve names are those found on the
Senmut circles, so that the identification is certain. It is this listing of monthly deities, combined with the fact
that an integral part of the Ramesseum upper panel is a strip divided into thirteen equal sections, in twelve of
which are given the months and the seasons, which makes the arrangement so important a clue to the type of year
depicted here. The essential elements are abstracted in Figure 19, where dotted lines join the individual month
rectangles to the offering scenes of the lowest panel, denoted in each case by the deity concerned. The names
added in the four parentheses are the variants from the Senmut ceiling. The most striking feature of the Rames-
seum ceiling, now clearly observable in Figure 19, is the central blank rectangle of the top strip directly above
and so obviously related to the cynocephalus on the djed-pillar, the well-known symbol of Thoth. What is here so
graphically represented is, I am convinced, the intercalary thirteenth month of the original lunar calendar with
its eponymous deity, Dhwty, whose feast, Dhwtyt, occurred in it.
225. The Ramesseum ceiling must, then, be regarded as a schematization of the original lunar calendar; and
indeed a more convincing representation is difficult to imagine.
226. THE NAMES OF THE MONTHS. We have yet to discuss the genesis of the civil year and after that the
transference of feasts from the lunar year to the civil. The proper place for such discussion is later. It is, how-
ever, necessary at this time to consider the names borne by the months, both lunar and civil, so that a path through
the complexities of the subject may be afforded the reader. It was long ago pointed out by Brugsch that the names
of months derived from important feasts celebrated in them.
9 0
I shall argue in this section that the Ebers calendar
and the Senmut ceiling, with confirmatory data from the Ramesseum ceiling, give us the list of the monthly festi-
vals of the original lunar calendar, that these names remained virtually unaltered at least through Ptolemaic times,
that after the creation of the civil year the lunar festivals found a corresponding place in it, but that in the latter
they were in the main superseded as eponymous feasts by others, some of which are the originals of the civil
month names found in Aramaic, Greek, and Coptic documents. All this is, then, merely preliminary to an investi-
gation into the lunar month names for their evidence on the nature of the first lunar year.
227. For the lunar calendar the point of departure must be the astronomical frieze of the outer hypostyle hall
of the temple of Edfu (Pls. IV- V). Here are depicted, after the decans
9 1
and the constellations of the southern
and northern skies and the planets, the fourteen deities of the fourteen days of the waxing moon, shown before a
staircase with fourteen steps, at the top of which is the wdit-eye, the symbol of full moon on the 15th day, with
Thoth adoring it. Following this are listed the deities of the thirty days of the lunar month, with each day and
each deity named, after which (and this is what is important for our purpose) come the personified months of the
year with their names. Since these months follow immediately upon the 30 lunar days, it is a quite justifiable
conclusion that they in turn are lunar and the year they make is a lunar year. Last of all appear three figures
43
oi.uchicago.edu
prt t
4
mw prt
2____________________________2__________4___3__2_ I
t I I!
I III I 1 1I
1I 1I1 /
1 II1I
1I 1 III I I I
I 1I I11
1 1 1III I / I1
tI III I / I
1II I I I ! 1
1II 11 I 1I I I
1 I I II!!
1t1II r I
t II I r I
II% I I I
I I I I 1
SI IIII
I I I I I I
II I I I
I IrM IS
m th '-P h 1t D w r- It H t-ti H s
1 I . 1 ., y 3- I pYI
(.sf bdt) k hr ki)1 (mnht) I (wp rnpt) I
I III
I I I I I I
FIG. 19.-The original lunar calendar as depicted in the Ramesseum astronomical ceiling.
FP
AP
oi.uchicago.edu
for the seasons.
228. Now if the names of the months of this lunar year be compared with those of the Ebers calendar and the
Senmut ceiling, it will be seen that they are virtually identical, except for the last month, which at Edfu is not
wp rnpt but R1-hr-ity, the same variant as is witnessed by the Ramesseum ceiling. Moreover, the eponymous
deities of the Ramesseum ceiling are in general the same as those of the Edfu frieze. We have, then, a well
attested and unchanging lunar year, so far as the names of its months are concerned, from the early 18th dynasty
down to the Ptolemies. This fact can be emphasized by reference to the names which the months of the civil year
bore at the same late time. These have already been listed in 22, but more of an explanation is desirable than
was there offered.
229. In the Persian period of Egyptian history and later the ordinary business documents and temple records
continued to record the months in dates after the earlier formula of I-IIII t, prt, and smw. These, however, as
Griffith pointed out,
9 3
were merely symbols, pronounced in each case by the name: Thoth, Phaophi, etc. This we
know from the use of these names in the contemporary Aramaic papyri and tablets and from the Greek forms in
the Ptolemaic papyri, as well as from the later Coptic. They were, like those of the lunar calendar, derived from
festivals in each month, and their hieroglyphic equivalents have been traced back as far as the 19th and 20th dynas-
ties.94 This is not to say that at that early time the normal dating by month and season was regarded only as a
symbol, as later, since we have such expressions as IIII prt 1 p-n Rnwtt, "IIII prt 1, the one of Renutet
(Pharmuthi),"
9 5
and II smw m p-n'Int, "II mw, being the one of (the feast of) the Valley (Payni).
" 9 6
The former
is from the first half of the 20th and the latter from the late 19th dynasty. There seems little room for doubt that
in the Empire the month and season continued to be read.
230. In Table 7 are presented for comparative purposes the names of the months of the lunar year and the
names of the months of the civil year, as attested in the 18th to 20th dynasty. Study of this table is fruitful and
illuminating. It is immediately apparent that the names of four lunar and four civil months are the same, those
of Ht-hr, k) hr k, Rnwtt, and Hnsw. This, to my mind, is a clear-cut indication that at some earlier time, nearer
the date of origin of the civil calendar, all the months of the civil year had borrowed their popular names from the
lunar year. As the centuries passed, however, and the civil year became more and more the important year in
TABLE 7
NAMES OF THE MONTHS
Lunar Calendar Month Names Civil Calendar Month Names
Intercalary Dhwtyt Early Late
I ht ty Dhwty Thoth
II " mnbt p-n'Ipt Phaophi
III " Ht-hr H t-hlr Athyr
IIII " k3 lhr k k 1 hr k3 Choiak
I prt sf bdt t3o C 3bt Tybi
II
III
IIII
I
II
III
IIII
p3 hnw Mwt*
" rkh wr
p-n mbr Mechir
" rkhl nd. p-n Imnlytp Phamenoth
" Rnwtt p-n Rnwtt Pharmuthi
mw nw p-n Unsw Pachons
" nt-hty p-n nt Payni
'Ipt .mt ipip Epiphi
" wp rnpt (mswt Rc -1hr-3ity)t Mesore
RC -hIr-bity wp rnpt
p3 mt n Hri
*The more usual early name; cf. Cerny, Ann. Serv., XLIII, 175.
t This is not yet attested as a month name in the 20th dynasty. The
reference in dernf is to the feast on I 3t 1 and not to IIII mw,
See 236 later.
$ The probable name for IIII mw in ostr. Brit. Mus. 5639a, 4.
45
oi.uchicago.edu
the life of the people, with new festivals having their places in it instead of in the older lunar year, it would not
be at all surprising that certain festival names should have given way to newer and more popular ones. The clear-
est example of this is perhaps that of p-n 'Imnlhtp, "the one of Amenhotep (the deified Amenhotep I of the 18th
dynasty)." This feast could not have come into being until after his death; but by the 20th dynasty it had become
so popular that it superseded as a month name its presumable predecessor, rkh nds. The archaizing inscriptions
of the Ptolemaic period, however, at times discarded the current names in favor of the earlier ones taken over
from the lunar months. We have already noted in one of the double datings discussed in chapter I (103) that the
civil month was given not as III mw but by the name 'Ip hmt-
4
. This can only be the old lunar name which was
later replaced by Ipip, Epiphi. In the same building inscriptions the month of I prt is given as sf bdt,
9 8
and this
must have been the predecessor of t3 c 'bt, Tybi. Besides these six names in common, the lunar intercalary feast
of Dhwty found a place in the first month of the civil year (181), along with ty, and eventually gave that month
its name. The variants for the last month of the year, both lunar and civil, we shall shortly treat of in some detail.
231. What names were borne by the months of the later lunar calendar, the one which ran concurrently with
the civil year, is not beyond conjecture. We have seen that a double date (84) gave the lunar equivalent as "the
6th lunar day of hb int," and that lb int could be nothing other than a variant of p-(n) int, Payni, the name of the
civil month II" smw. In all probability, then, when the months of the later lunar calendar were referred to by name,
those names were the ones borne by the civil months, and not those of the original lunar calendar. This, of course,
is exactly what we should expect.
232. Having cleared the ground, we are now ready to search into the month names of the original lunar calendar
for evidence of its structure. If the rising of Sothis did control this calendar, we may expect to find some appro-
priate indications, natural and seasonal, in the names of the months. In actual fact, four names do afford us such
indications. They are (5) f bdt, (6) rkh, (7) rkh, and (8) Rnwtt, the four months of the season prt; these are,
respectively, the month of "Swelling of the Emmer,"
9 9
the two months of "Fire," usually differentiated by "Great"
and "Small,"100 and the month of "Harvest," in the person of the harvest goddess.
233. In 3100 B.C., approximately the date of the union of the Two Lands, Sothis rose heliacally about June 20
greg. The first month of a lunar year regulated according to our thesis would begin within the approximate limits
of July 1 - 30. "Swelling of the Emmer" would then fall approximately in November-December, the two "Fire"
months in December-February, and "Harvest" in February-March. The exact meaning of Af bdt is not clear, but
it undoubtedly had some reference to growth. November-December would be a not inappropriate time, since plant-
ing of seed goes on steadily through September and October as the flood water recedes from the land. "Harvest"
in February-March would also fit quite well.101 The two months of "Fire" require more detailed consideration.
234. At Cairo the mean maximum temperature is 690 in December, 660 in January, 72

in February, 74

in
March. The mean minimum for these four months is 50
. At Luxor the mean maximum is 760 in December, 740

in January, 780 in February, and 85

in March, with a mean minimum of 500 for all four months. 102Clearly,
December-February is the time when some degree of artificial heat would be welcome. Now all occurrences of
rkl in the festival lists of the Old Kingdom described above (176-81) exhibit the determinative of the brazier
with a flame rising from it (a) -artificial heat. The conclusion, however, that the principal feasts of two months
were concerned with such a phenomenon might be considered somewhat rash, did not an exact parallel from Meso-
potamia firmly establish the practice.
1 0 3
There the month later known as Arahsamna was originally called warah
kiniinim, "the month of the brazier or hearth-fire"; and indeed the characteristic ceremonies of Arahsamna,
Kislimu, and Tebetu (November-January) were based upon the kinfinu, the hearth fire.
1 0 4
The month of warah
kinilnim has also been found in the calendar of Mari -a further indication of how natural the use of such a name
was to the ancient Near East.
235. Sethe, who had a theory that originally the year began with the winter solstice and the month miwt rC,
"Birth of Re," a variant name, as we have seen (213), for the month wp rnpt, considered the rklh-months to be
those of solar heat in the summer; but, in order to fit in ~f bdt and Rnwtt properly before and after them, he had
to make them later additions to the month names, after the year had shifted forward.
1 0 6
Borchardt rightly objected
that there was very little basis for such a theory;107 and Sethe himself pointed out that there is no evidence at all
in Egyptian records to demonstrate the four points of the sun-a clear-cut argument against recognizing the winter
46
oi.uchicago.edu
solstice in mswt Rc.
1 0 8
236. If one thus rejects.Sethe's explanation of the name mswt RC, he is in duty bound to offer an alternative.
This I now propose to do. The following facts are at our disposal. The primary meaning of wp rnpt was the helia-
cal rising of Sothis. The festival celebrating this event gave its name to the last lunar month of the year. A vari-
ant name for the same month derived from the festival of the birth of Re. Both these names and probably one
other, p 3mt n Hr, "the going-forth of Horus," were taken over as designations of the last civil month. A second-
ary application of the name wp rnpt was to the first day of the civil year, I i t 1. This day was also known as
mswt RC -hr- hty at least as early as the 20th dynasty
1 0 9
and as ms(wt) tn, "the birth of the sun disk," and ms(wt)
110
RC in the Ptolemaic period. The two names were frequently combined at Dendera when the feast celebrated
,,111
on I 3ht 1 was referred to as hb Rc m wp rnpt, "the feast of Re as the opener of the year. Gardiner, seeing
in wp rnpt a solar feast, held it to commemorate "the moment when the sun-god, in his first act of rising, opened
the succession of months and years, as the originator of which he is so often eulogized. But the first rising of Re
was also the instant of his 'birth' (Mesore), the occasion of his earliest 'going forth' (p, mt n Hr).
237. Gardiner's statement should be applied, I believe, not to the day of the secondary application of wp rnpt
but to the day of its primary meaning, the rising of Sothis. It was on such a day, according to the tradition pre-
served by two late classical writers, that the world came into existence. According to Porphyry (de antro
nympharum 24),
for the Egyptians the beginning of the year
ABiyyroes be Apx' irov oby' bbpoxboo, ws 'Pwgaiots, aXMa Kaprt-
is not Aquarius, as'for the Romans, but
Vos. rps Yd&p rTC KapKtLw f , 8s, qv KvvoS arTEpa "EXXy
1
rves sa. Cancer. For near Cancer is Sothis, which
vov7via 5' atiros acOews dvaroX , yevEaEws KarapXovaa rfs Els T r, the Greeks call the dog-star. Their new
moon is the rising of Sothis, which is the
Ka'ov.
/ beginning of generation in the world;
and Solinus (Polyhistor 32. 12-13) states that
quod tempus sacerdotes natalem mundi indi- this time (the rising of Sirius) the priests
caverunt, id est inter XIII k. Aug. et XI. have decided to be the birthday of the world,
that is, the time between the 13th and 11th
days before the Kalends of August (July 20-22).
To the ancient Egyptian who used a lunar year based on the rising of Sothis, any other day than that one for the
creation of the universe would have been unthinkable, for that event determined the months and the seasons in
their proper succession. How could the world have begun on a day in the season of prt, for example? There
would have been no inundation, and the seed would not have been sown. But shortly after Sothis rose heliacally
for the first time, and the world came into being, the sun-god Re also rose for the first time. That first day of
wp rnpt was also, in very truth, the day of miwt Re. So easily, then, could the one day have two names and the
last month of the lunar year and the civil year have two names. So easily also, when the one term of wp rnpt was
taken over and applied to the first day of the civil year, could the second term mswt Rc have been taken along with
it and applied to the same day.
238. THE FEAST OF MIN IN THE PTOLEMAIC PERIOD. In lines 13 and 14 of Cairo Dem. Pap. 30801 (90)
occurs the significant entry: "Horsiese, son of Hetpesobk, for the service from III prt 19 to IIII prt 17 [of the
third phyle, for the grain-rations] and the food of the priests for the feast of Min, and the extra expense of the
temple, wheat 10 (artabas)."
At Gebelein, then, in the 26th year of Ptolemy VIII Euergetes II (144 B.C.) or, as a remote possibility (97), in
his first year (169 B.C.), there was a festival of Min which took place on some day or days of the lunar month be-
ginning III prt 19. The calendar of Esna has a procession of Min-Amon on I Emw 1. Whether this date is lunar or
civil is uncertain, but not significant, since in A.D. 175-76 when the Esna calendar was compiled (see the following
section) the original lunar year and the civil year were again in complete harmony. The Medinet Habu calendar,
however, as we haver, as we have already see204), dates the feast of Min (204), dates the feast of Mto pntyw of mw, and, since at this time the civil
year and the natural year were almost in harmony, it is a logical deduction that the lunar month whose first day
began the feast was likewise I mw. We have now to see whether the original lunar calendar, which provided a
47
oi.uchicago.edu
natural calendar for such harvest festivals, can explain the equation III prt 19 (civ.)
=
I mw 1 (lunar).
239. In 145 B.C. wp rnpt took place on July 19 or 20, II mw 25 or 26 (civ.), year 12 of the cycle (49-64) be-
ginning 157 B.C. In this year the first lunar month after II 6mw 25 began on II 6mw 28. Since this date is within
eleven days of wp rnpt, according to theory it must have been the intercalary month Dhwtyt. By continuing with
the cycle dates, we arrive at the following table:
Civil Year Lunar Year
II mw 25 wp rnpt July 19, 145 B.C.
II " 28 July 22 Intercalary Dlhwtyt
III " 28 Aug. 21 I .t thy
IIII " 27 Sep. 19 II " mnbt
I ht 21 Oct. 18 III " Ht-hr
II " 21 Nov. 17 IIII " k hr k
III " 20 Dec. 16 I prt 9f bdt
IIII " 20 Jan. 15, 144 B.C. II " rkl wr
I prt 19 Feb. 13 III " rkh nd.
II " 19 Mar. 15 IIII " Rnwtt
III " 19 Apr. 14 I mw ndw
240. We might have placed Q.E.D. after the last line of the tabulation, since the equation comes out exactly.
This is, to my mind, not only a convincing proof of the correctness of the proposed theory of the lunar year but
also a smashing refutation of the need for a fixed Sothic year postulated by Sethe and Weill. Natural feasts were
governed by the natural lunar year from the beginning of Egyptian calendar-making right down to the last days of
Egypt under the Greeks and the Romans. Nor is this affirmation vitiated by the statement of the decree of Canopus,
according to which it was desired that the feasts which were celebrated in winter should not come to be celebrated
in the summer (line 21). That this had nothing to do with the natural year but was simply a desire to retain the
status quo of the civil year is apparent from the means taken to attain the end. Since the rising of Sothis took
place on II 9mw 1, the civil year was at variance with the natural year by a full three months. Instead of correct-
ing the calendar by advancing it three months, which would have been necessary had it been desired to fit the civil
year to nature, and thereafter having a sixth epagomenal day every four years, all that was proposed was to have
a sixth epagomenal day every four years and freeze the status quo. Such a decree inferentially argues for the
existence of a year which actually did control the natural feasts, so that the attempt would be merely to keep to
their present seasons the feasts of the civil year which really had no concern with nature.
241. It was mentioned above that there was a remote possibility that the temple account was to be placed in the
first year of Ptolemy VIII, 170 B.C. The tabulation for that year also would result in verification of the equation
and of the proposed lunar calendar. Wp rnpt fell on II mw 18, and again, since the following lunar month began
within eleven days, on II gmw 28, it would be intercalary.
Civil Year
II mw 18 wp rnpt
II " 28
III " 28
1111 " 27
I t 21
II " 21
III " 20
IIII " 20
I prt 19
II " 19
III " 19
242. THE FEAST OF RENUTET IN THE
Lunar Year
July 19, 170 B.C.
July 29 Intercalary Dhwtyt
Aug. 28 I 3het thy
Sep. 26 II " mnht
Oct. 25 III " Ht-hr
Nov. 24 IIII " ko khr ki
Dec. 23 I prt Af bdt
Jan. 22, 169 B.C. II " rkh wr
Feb. 20 III " rkh nd
Mar. 21 1111 " Rnwtt
Apr. 20 I mw Unsw
ESNA CALENDAR. It was stated above (200) that there were two
48
oi.uchicago.edu
feasts of Renutet listed in the calendar of the temple of Esna. One fell on the usual date of I mw 1, the other on
I ,mw 25. The explanation was suggested that there was both a fixed and a movable feast of Renutet, just as there
was a fixed and a movable wig-feast in the Middle Kingdom. If this suggestion be correct, then the movable feast
ought to be one celebrated according to the natural lunar calendar. In order to test this, we have first of all to
assign a date to the compilation of the Esna calendar.
243. In this calender
1 1 4
there are listed three feasts of wp rnpt: (1) on I 3ht 1 1.1); (2) on I Tht 9: ,
"the feast of wp rnpt, so called by the ancestors," (1. 1); (3) on II mw 26 (1. 15).
244. The Esna calendar is not dated to any king, but the royal names in the reliefs which flank it suggest
that it must fall in the period from Domitian (A.D. 81-96) to Caracalla (A.D. 211-17). The natural assumption is
that the wp rnpt of I ibt 9 is the rising of Sothis. With an arcus visionis of 90 and Memphis as the point of obser-
vation, Sothis would rise on I 3ht 9 in the years A.D. 173-76, in the reign of Marcus Aurelius. With Esna as the
observation point, the years would be 190-93, all but the last being in the reign of Commodus. Both of these kings
had a part in the decoration of the hypostyle hall, and the calendar could thus be ascribed to either.
245. In line 17 of the calendar, however, we have the significant entry: "Day 20 (of IIII mw). Feast of the 29th
lunar day (hb ChC ... )." Why this particular lunar date was incorporated into the calendar we shall never know, but
it is of prime importance. We learn from it that, in the year from which calendar dates were drawn, a lunar month
began on III mw 22. At this time the 25-year cycle must still have been in operation. III .mw 22 falls in the
seventh cycle year, and this can only be A.D. 175/76.116 Thus lunar date and Sothic date agree on this year in
the reign of Marcus Aurelius as the compilation date of the Esna calendar. It is significant, moreover, that
Memphis was still the traditional observation point for Sothis for all Egypt.
1 1 7
246. To make conclusive the identification of the wp rnpt of the ancients on I t 9 as the rising of Sothis, we
must offer some explanation for the wp rnpt of II smw 26. We have seen that in the decree of Canopus the birth-
day of the king was called wp rnpt. This would not do for Marcus Aurelius, who was born on April 26. In A.D.
161, the year of his accession, II smw 26 was May 5 and in A.D. 121, when he was born, it was May 15. It is my
belief that what is commemorated here is rather the accession of the emperor to the throne of Egypt.118 At Rome
he acceded on March 7. A period of 59 days (to May 5) seems sufficient for the news to reach the prefect of
Egypt, who would then order a holiday and a period of celebration.
1 1 9
247. Returning now to the lunar date of III smw 22 and the seventh cycle year, we proceed to tabulate that year
just as we did above for the feast of Min, with the following result:
I ht 9 wp rnpt July 19, A.D. 175
I " 26 Aug. 5 I 3ht thy
II " 26 Sep. 4 II " mnbt
III " 25 Oct. 3 III " Ht-lhr
1111 " 25 Nov. 2 1111 " k? hr k'
I prt 24 Dec. 1 I prt 9fbdt
II " 24 Dec. 31 II " rklh wr
III " 24 Jan. 30, A.D. 176 III " rkh nds
IIII " 23 Feb. 28 IIII " Rnwtt
I Emw 22 Mar.28 I Emw Undw
I " 25 lb Rnwtt I " 4 (prt sm)
248. There is nothing at all suggestive about a feast of Renutet occurring on prt Am of the lunar month of
Indw, and we may reject with finality any connection between the date of the feast and the above lunar year. Let
us, however, continue the tabulation into the following lunar year.
II .mw 22 Apr. 27, A.D. 176 II mw nt-hty
III " 22 May 27 III " "Ipt Imt
IIII " 21 June 25 1111 " wp rnpt
49
oi.uchicago.edu
I ht 9 wp rnpt July 18
I " 15 July 24 Intercalary Dhwtyt
II " 15 Aug. 23 I iht tb
III " 14 Sep. 21 II " mnht
IIII " 14 Oct. 21 III " Ht-hr
I prt 13 Nov. 19 IIII " k- hr k'
II " 13 Dec. 19 I prt fbdt
III " 13 Jan. 18, A.D. 177 II " rkhwr
IIII " 12 Feb. 16 III " rkh nds
I mw 11 Mar. 17 IIII " Rnwtt
I " 25 hb Rnwtt IIII " 15 (smdt)
249. We need continue no further with the tabulation. A feast of Renutet falling on smdt, the day of full moon,
in her own lunar month is so correct and appropriate that its identification cannot be denied. Nor can doubt be
cast upon it by the fact that it is in the year after the one from which the dates for the calendar were apparently
compiled. We have no reason at all to suppose that the calendar was composed at the very end of the civil year,
with all its dates taken from the temple records of the preceding twelve months. It could have been done at any
time in the year, and that is what must have happened in this case. Some time after I ,mw 25 of A.D. 177, but
before IIII mw 20, a temple scribe compiled a festival calendar for the temple of Esna, using the temple records
for the preceding twelve months. He went as far back as, and no doubt somewhat before, IIII mw 20, because that
date had to fall in A.D. 176, being the 29th day of the lunar month beginning III smw 22. This may perhaps be seen
more easily by reference to the above tabulations. One has only to begin with the end of the table and count back
twelve months to find that IIII mw 20, the lunar feast of the 29th day, falls well within the twelve-month range of
calendar compilation.
250. There was a feast of Renutet celebrated at Edfu on I prt 7 which cannot easily be accounted for on the same
120
basis as that of I mw 25 at Esna. For the equation I prt 7 (civ.) =- IIII prt 15 (lunar) to be correct, the Edfu
calendar would have to date from some year of the fourth or late third century B.C.;
1 2 1
but the foundation of the
122
temple did not take place until 237 B.C. It is certainly true that there was a temple at Edfu prior to the present
structure, and it cannot be entirely excluded as a possibility that when the calendar was carved on the walls an
earlier version was utilized. But, since this seems rather unlikely, an alternative explanation is to be preferred.
Only two possibilities have occurred to me, the first of which is that the feast was not of Renutet, the goddess of
the harvest, but of Renutet, the goddess of clothing.
1 2 3
A feast to the latter could presumably take place any time
in the year. The second possibility envisages a rather daring textual emendation.
251. The Myth of Horus, in one of its episodes, relates all that happened on a certain day, I prt 7. Thereafter,
Thoth decided, I prt 7 should be called hb hn, "the feast of rowing.",
1 2 4
This date is referred to again, in
company with I tht 1 and II prt 21 and 24, as an important festival of Horus.125 It would be very strange,
then, if such an important feast found no mention in the temple calendar; yet the only entry under I prt 7
is hb Rnnt.
1 2
I suggest, with diffidence, that an original hieratic writing of ,12 hb hn,
with the n sign written rather badly, was mistaken by another scribe for and then filled out to
^""" % . Of the two possibilities, I myself incline toward the latter, even though it does involve an emendation.
Otherwise one would still have to account for the omission of the hn-feast from the temple calendar.
252. CONCLUSION. We have in this chapter traced the first lunar calendar throughout the long course of Egyp-
tian history. We have seen that it affords a satisfactory explanation of the most diverse phenomena: the calendar
of feasts in the Old Kingdom mastabas, the Middle Kingdom dates of the movable wg-feast, the temple year at
Illahun, the Ebers calendar, the astronomical ceilings of Senmut and the Ramesseum, the names of the months, and
the feasts of Min and Renutet in the latest periods. It has met every test that can be brought against it at this time.
253. Since the original lunar calendar must then be counted a certainty, we are confronted with the situation that
in the later period there were three calendars in use. We have still before us the problems of the origin of the
civil calendar and of why and when the later lunar calendar was introduced. To these problems we may now turn.
50
oi.uchicago.edu
CHAPTER IV
THE CIVIL CALENDAR
254. THE PROBLEM BEFORE 1938. The general theory of the civil calendar is so well known that it is un-
necessary to deal with it in great detail. As expounded by Eduard Meyer in 19041 it involves an early determina-
tion by the ancient Egyptians of the length of the solar year as 365 days, a division of this year into three seasons
of four months each and five epagomenal days, and the recognition of the heliacal rising of Sothis as marking, at
calendar inauguration, the first day of the year. Since, however, the rising of Sothis did not take place every 365
days exactly, but rather every 365 1/4 days, after four years the civil year would begin on the day before the rising
of Sothis and, after eight years, two days before that event. In 1,460 Julian years (the Sothic period) the civil cal-
endar would move forward through every day in turn until it had completed the circuit and once again its first day
fell on the day of Sothis' rising. One of these coincidences, according to Meyer's reckoning, took place on July
19, 2.781 to 2778 B.C.; but, since the civil calendar was demonstrably in use before that date and since it must
have been inaugurated on a coincidence day, he went back 1,460 years to 4241 B.C. and claimed July 19 of that year
as the first certain date in history. This was indorsed by Breasted and became the cornerstone of Egyptian chro-
nology.
2
255. As far as the origin of the civil calendar was concerned, little or no attention was paid to lunar reckoning.
We have remarked above (147-48) that Meyer believed that some sort of lunar year was in use before the intro-
duction of the civil year but considered the civil year to be a purely artificial creation, since neither month nor
season nor even year corresponded to any natural period; while Sethe was of the opinion that the lunar year and
the civil year were parallel in'development, since the former had no proper place in the "Entwicklungsgeschichte
des aeg. Jahres." He suggested rather a year of 360 days as the preliminary stage to the 365-day year. We might,
therefore, sum up the commonly accepted theory of the civil year as follows:
1. While lunar reckoning or a lunar calendar existed before the civil calendar, the latter had no con-
nection with it but was an independent creation.
2. The year had been "astronomically" determined as being but 365 days long.
3. The first day of the year when inaugurated was the day of the heliacal rising of Sothis.
4. Despite the fact that the first day of the year began to fall more and more in advance of Sothis'
rising, no adjustment or correction was ever made.
256. THE PROBLEM AFTER 1938. In that year Otto Neugebauer published an important and provocative
study of the Egyptian calendar.
3
His paper divides naturally into two sections. In the first he attacks with vigor
the conception of the Sothic period as an instrument for determining, as Meyer did, the oldest certain date in
history to be July 19, 4241 B.C. He demonstrates that there could not have existed at that time any theoretical
astronomy, as writing and mathematics did not yet have any being and the cultural level of the people must have
been quite low. He shows that the body of astronomers credited by Borchardt with the revolutionary installation
of the 365-day year can have existed only in the latter's imagination. He emphasizes the contradiction between
the revolutionary character of a 365-day year based on Sothis and the failure of the proponents of this new calen-
dar to adjust it to Sothis when after but eight years the year began two days before the rising of Sothis. He con-
cludes that while the Egyptians had two conceptions of the year-(1) a period of 365 days; (2) the interval between
two risings of Sothis -in the beginning these had nothing to do with each other.
257. In the second section of his paper Neugebauer presents his own theory of the origin of the 365-day year,
entirely apart from Sothis. He shows that an averaging of the intervals between inundations over a period which
need not be greater than fifty years would inevitably result in an interval of 365 days. If, then, this "Nile" calen-
dar were adopted in a year when the inundation was normal, because of the great variability of the inundation it
would be some centuries before the calendar seasons no longer coincided with the natural seasons. A new phe-
nomenon would then be picked as expressing more clearly than the calendar the incipient inundation. This was
the rising of Sothis.
51
oi.uchicago.edu
258. When he first announced his theory, Neugebauer was inclined to place the introduction of the 365-day
year in the centuries around 4200 B.C. Recently, however, as the result of a study of his theory by Scharff,
4
he
would place it in the centuries around 2800 B.C.
5
259. In my opinion weighty objections can be brought against either time of introduction. Neugebauer has him-
self declared that "the only condition for the creation of the schematic calendar is a sufficiently well organized
and developed economic life,"
6
a doubtful claim for the Egypt of ca. 4200 B.C. Moreover, as Winlock points out,
"when one Nile year might be only 335 days long and another as much as 415, it is a question whether primitive
man would ever, unaided, have arrived at the conception of an average Nile year or would have known how to calcu-
late it, had he thought of it."
7
However, since Neugebauer no longer considers ca. 4200 B.C. as the probable date
of introduction, we may limit ourselves to a consideration of a date ca. 2800 B.C.
260. It is unquestionably true that by this time Egypt possessed a well organized and developed economic life,
as well as writing and mathematics. But Egypt also possessed at this time, as has been shown in the previous
chapter, a "Nile" lunar calendar based on Sothis. Neugebauer's theory fails to take into account the already exist-
ing calendaric situation. Obviously there would be no point to averaging the intervals between inundations in order
to arrive at a Nile year, when all the time there was present and in use a lunar Nile Year. But if we reject the
average interval between inundations as the ultimate basis for the 365-day year, we must, I believe, hold fast to
the other point which Neugebauer has so forcefully demonstrated. The civil calendar of 365 days was not tied to
Sothis at its introduction but was tied rather to some yearly occurrence which was variable, so that the gradual
shift forward of the civil calendar would not be immediately apparent.
261. It is this point which forces one to reject the recent theory of the origin of the civil calendar proposed
by Winlock. After detailing the steps through which primitive man passed toward a recognition of Sothis as the
herald of the inundation, Winlock suggests that around the time of Menes the Egyptians were beginning their year
with the reappearance of Sothis. This year no longer consisted of lunar months but rather of 30-day months,
divided into the three seasons. Over and above these months were a few extra days every year, usually five but
at times six or rarely four because of faulty observation of the heliacal rising of the star. In any event all varia-
bility was confined to the extra days. Eventually, around 2773 B.C., probably in the reign of Djoser, observation
for beginning the year was dropped, as "the experience of centuries by now had seemed to show that the year
should contain 365 days, and this definite figure was adopted for administrative purposes."
8
262. However, a century of recording the extra days of the year would give a total not of 500 but rather of 525,
a clear indication that the year could not be exactly 365 days long; and eight years after the adoption of the definite
year it cannot fail to have struck those responsible for its use that the first day of the year was already two days
in advance of the rising of Sothis and that inevitably the calendaric seasons would be at variance with nature. The
correction of this situation by an extra epagomenal day every four years is so simple and obvious that failure to
make such a correction argues that the situation never existed in reality and that the 365-day year was never tied
to Sothis.
263. There are other objections to Winlock's theory. He would explain tpy rnpt in the mastaba lists as being
the New Year's Day invented for the calendar after it had become separated from nature. We have seen, however,
that these lists were arranged chronologically, which would not be the case were tpy rnpt a day moving farther and
farther forward from the rising of Sothis, wp rnpt. Moreover, his theory puts the epagomenal days after the twelve
months; our only Old Kingdom evidence on the point is to the effect that they came before the months.
9
Winlock
fails too to take into account the evidence of the existence of a lunar calendar of a well developed type.
2 64. THE PROBLEM AT PRESENT. There are, it would now appear, certain fundamentals which cannot be
ignored in considering the problem of the origin of the civil calendar. These have all been brought out in the pre-
ceding pages, but it will be useful to summarize them.
1. The presence of a well regulated and working lunar calendar must not be ignored.
2. The civil calendar must at its introduction have been tied not to Sothis but rather to some event which
was variable in itself, so that the forward shift of the year would not be immediately apparent.
3. The proposed theory of the civil calendar must include an explanation of why the later lunar calendar,
tied to the civil calendar, was inaugurated.
52
oi.uchicago.edu
THE CIVIL CALENDAR
265. THE PROPOSED SOLUTION. During the protodynastic period the only calendar in use was the lunar
calendar already described. By this time Egypt had become a well organized kingdom, and the economic dis-
advantages of a lunar year of now twelve months, now thirteen, all of which began by observation, must have im-
pressed themselves upon the government. In an effort to alleviate the situation and to provide a simple and easily
workable instrument for the measurement of time, they hit upon the idea of a schematic lunar year or, as it might
be termed, an averaged lunar year. There are two ways by which the length of this schematic lunar year might
have been determined as being 365 days. On the one hand, since the current calendar was based on Sothis, one
might simply have counted, for one or two years, the number of days between successive heliacal risings of the
star. On the other hand, the lunar year itself might have been averaged. It would have been little trouble to refer
to the various records of one kind or another from which data on the number of days in recent calendar years
might be derived. Over a period of 25 lunar years something like the following might have been arrived at:
1 0
1. 354 9. 355 17. 384
2. 354 10. 354 18. 355
3. 384 11. 384 19. 383
4. 354 12. 354 20. 354
5. 354 13. 354 21. 355
6. 384 14. 384 22. 383
7. 355 15. 354 23. 355
8. 384 16. 355 24. 354
25. 384
Now if these figures be averaged in cumulative fashion from year to year, it will be soon apparent that the mean
must be 365. In the following table the averages are limited to those years wherein the corrective factor of the
intercalary month occurs, and it can be seen that already after eleven years the round number of 365 days must
have impressed itself upon the ancient Egyptian with some force.
1 1
3. 364 11. 365.09 19. 365.21
6. 364 14. 364.85 22. 365.04
8. 365.37 17. 364.76 25. 364.96
266. Thus easily, by either method, could the Egyptian of the protodynastic period have arrived at 365 days
as the proper length for a schematic lunar year which could be adopted for administrative and economic purposes.
After the analogy of the ordinary lunar year the schematic year would be divided into twelve months and three
seasons, each month having 30 days for simplicity and regularity.
1 2
The extra five days of the schematic year
were regarded as an abbreviated intercalary month and were placed before the year, just as the intercalary month
headed the lunar year whenever it occurred.
267. Now as the new year was after all only a schematic lunar year, it must have been planned to have it run
concurrently with and as far as possible in concord with the old lunar calendar, which was by no means to be
abandoned. Just what were the actual circumstances of its introduction we shall probably never know. It is pos-
sible, at least, that it was introduced in a year with an intercalary month and that the five epagomenal days were
concurrent with the last five days of that month, so that the first day of both the old lunar year and the new sche-
matic year fell on tpy rnpt, the first of thy; or it may have been handled without regard for the epagomenal days,
with the new year, at its installation, simply having its first day coincide with tpy rnpt.
1 3
268. If we cannot be certain of the circumstances of the introduction, we can at least set a range in time with-
in which it took place. According to the rule for intercalation in the lunar calendar, when the first month after
wp rnpt began within eleven days of that event it was intercalary. Tpy rnpt might then be as close to wp rnpt as
the twelfth day following or, with a full intercalary month beginning on the eleventh day following, as far away as
forty-one days. It is almost a certainty that the first day of the first month of the civil year, as we shall hence-
forth term the schematic lunar year, also fell within these limits. As this first day had come to be the date of
the rising of Sothis in ca. 2773 B.C., the civil calendar must have been introduced between ca. 2937 and ca. 2821
B.C.,14 with the probability that it was in the direction of the former rather than the latter date.
53
oi.uchicago.edu
269. Now an important feature of the new calendar was that complete agreement between it and the lunar calen-
dar was not expected, as the latter was so variable. Consequently the fact that the new calendar moved forward a
day every four years would be unnoticeable. The change would be so imperceptible that even after half a century
there would still appear to be quite good general agreement between the civil and lunar years (see Figure 20).
And all this time, of course, the civil calendar will have been proving its worth. All secular matters would no
doubt be regulated by it, while to the lunar calendar would be left the determination of feasts and the regulation of
temple service. The two years were distinct but intrinsically the same, the second, the civil year, being merely
an aspect of the first.
Day 1 of Day 1 of
civil year civil year
wp rn t after 50 years at installation
L -wp rnpt Dhwtyt thy
LEpj Month I C
L pt bmt L wp rnpt_ th
----------------- _-------------i----------------s
Month XI Month XII Ep. Month I C
L---- - -)---------t
Month
XI
Month
XII
Ep.
Month
I
C L wp rnpt twtyt thy
Month XI- -L Month XII--_-_---p-Month -
L ..... [. wp rnpt A thy p[.!.
Month XI Month XII _L -Month I C
FIG. 20.--Concurrence of lunar and civil years at installation of
latter and shift of civil year after fifty years.
270. After about two centuries, however, it can no longer have escaped attention that there was trouble in this
dualistic calendar set-up. No longer did the first lunar month and the first civil month touch, even for a day. How
long this situation kept worsening, we cannot guess; but when the decision was made that something would have to
be done, it was found that there were but two ways to redress it. The first would be to insert the requisite number
of days at the end of the civil year in order to force it back to concurrence with the lunar year as at its introduc-
tion. This action we may suppose to have been rejected with horror by officialdom. No doubt by this time the
fiscal year was firmly established as 360 days
1 5
and was not to be added to or diminished. It may very well have
been at this time, if we accept a late tradition, that the king swore never to tamper with the year-an oath repeated
at coronation by his successors to the end of Egyptian history.16
271. Now it must be remembered that the civil year was, after all, merely a schematic lunar year, one which
had no real being when separated from its natural counterpart. Obviously the lunar calendar itself, being controlled
by Sothis, could not be moved from its place in order to continue supporting the civil year by furnishing body and
substance to its artificiality. It would appear that some brilliant Egyptian solved the difficulty into which the calen-
dars had gotten by the creation of a special lunar year, whose sole purpose would be to provide for the civil year
the same sort of dualistic setup which had obtained when the civil year was first inaugurated. The months of this
new lunar year were to be so regulated that they would maintain their general agreement with their schematic
equivalents. In this fashion the original lunar calendar would continue on independently as before, while the later
lunar calendar and the civil calendar, the dual year, would be free to progress forward through the seasons.
272. Satisfactory as this explanation of the genesis of first the civil year and later the second lunar calendar
might be considered, the fact must be granted that it would continue to be merely a hypothesis were not the essen-
tially dual character of the year brought out unmistakably in the monuments.
273. THE FIFTY-NINE DIVINITIES OF THE DUAL YEAR. There are two series of divinities connected
with the year in Egypt, one consisting of the thirty-six decans, the earliest extant lists of which occur on the so-
called "diagonal calendars" of the Middle Kingdom, the other consisting of fifty-nine deities, many of whom bear
54
oi.uchicago.edu
THE CIVIL CALENDAR
decanal names, the earliest known list of which appears on the base of a statuette of Mut of the 22d dynasty. The
explanation of the thirty-six decans is superficially simple, as they were the thirty-six stars or constellations
which rose during the twelve hours of the night throughout the thirty-six decades of the year, each month of the
civil year being divided into three parts of ten days each; but the list of fifty-nine deities has so far resisted satis-
factory analysis.
274. The fifty-nine have both names and representations. Those of the Greco-Roman period, from the temples
of Dender, Edfu (P1. IV), and Esna, are conveniently tabulated in Brugsch, Thesaurus, pages 18 - 23. The earliest
list, representations only, is found on Cairo statuette 38924.17 Here and in the menat-fragment which we shall
shortly discuss the total number of representations is clearly fifty-nine. In the later lists there were frequent
omissions, so that only Dendera, with fifty-seven, approaches the correct total. Here the missing two are the
figures of Set, for the third epagomenal day, and a companion deity.
275. Study of the makeup of the list of the fifty-nine deities reveals a simple scheme. The first forty-eight
names and representations separate easily into groups of four. The first figure in each such group is almost
always a seated female with the head of a lioness, the second figure is almost always a standing male with the
head of a lion, the third figure varies but is predominantly reptilian, while the fourth and last figure is always a
serpent, usually standing on its tail. When grouped in fours, the first, third, and fourth prove to have ordinary
decanal names, while the second name in each group is an intrusion, in some cases an out-of-place decan, in other
cases newly coined.
1 8
276. In other words, the forty-eight names are merely those of the ordinary thirty-six decans expanded by the
insertion of one new name in every group of three decans. At the same time it must be pointed out that this was
quite clearly not a work of the Ptolemaic period, since at Dendera the normal list of thirty-six decans and the list
underlying the first forty-eight names are not exactly the same. This may be seen most easily by referring to
Brugsch, Thesaurus, pages 147 - 52, where the list I is abstracted from the forty-eight names and the list L is the
normal thirty-six. In twenty-six names there is agreement, in the remaining ten disagreement. Moreover, when
representations of the normal thirty-six decans occur they are not at all similar to the other depictions but are
instead mainly human figures, with various animal, bird, or human heads, who may be in barks.
277. The remaining eleven divinities begin with a serpent and then alternate the deities of the five epagomenal
days with other serpents or human figures. The statuette Cairo 38924, for example, has the following:
(49) A long serpent, reversed, on a support.
(50) The mummy of Osiris standing.
(51) A serpent erect.
(52) A hawk-headed Horus.
(53) A serpent erect.
(54) Set standing.
(55) Crocodile-headed deity standing and offering two jars.
(56) Isis standing.
(57) Man standing and offering two jars.
(58) Nephthys standing.
(59) Man standing and offering two jars.
278. The breakup of the fifty-nine deities into twelve groups of four, followed by one of eleven, which we have
established by analysis of figures and names, is confirmed strikingly by the preserved lower portion of a glazed
clay menat of the late period, now in the Cairo Museum collection (P1. VI A). There are three rows of deities
preserved. In the first row are recognizable the figures of Nos. 33, 34, 35, 36 on one side of a central bar and on
the other the lower parts of what must be Nos. 37, 38, 39, and 40. In the second row are found Nos. 41, 42, 43, 44,
the bar, and Nos. 45, 46, 47, and 48. In the last row are deities Nos. 49 to 59 inclusive, with some individual varia-
tions from the statuette list. Beyond doubt the object had once included all fifty-nine figures, with the first forty-
eight arranged in six rows, with two groups of four to each row.
279. Daressy is the only one to my knowledge who has attempted any explanation of the fifty-nine divinities.
He came to the conclusion that they should be divided into two groups of forty-nine and ten.22 On the assumption
55
oi.uchicago.edu
that each one of the forty-nine figures represented a week of 7 1/4 days and the remaining ten figures represented
1 day each, he arrived at a total of 365 1/4 days for the year. This explanation was ingenious but palpably incor-
rect. Not only is a week of 7 1/4 days absurd, but we have seen that the correct division is into groups of forty-
eight and eleven. Moreover, the idea that the Egyptians, even by the 22d dynasty, had a comprehension of the
length of the year as 365 1/4 days and a desire to portray that fact symbolically must be rejected.
280. But what, then, were these deities intended to represent? It is my belief that what we have is a depiction
of the dual character of the concept "year" as a combination of both the lunar and the civil year. The lunar month
divides easily and naturally into four parts: (1) days 1 - 7, from the month's beginning to and including the first
quarter; (2) days 8 - 15, from first quarter to and including full moon; (3) days 16 - 22, from full moon to third
quarter; (4) days 23 - 29/30, from third quarter to the end of the month
2 3
If we assume a presiding deity for each
one of these parts, we should have a total of forty-eight for the twelve months of the normal lunar year. Since the
remaining eleven deities of the second group include the five deities of the epagomenal days, it is an easy conclu-
sion that each of the eleven represents one day. Eleven epagomenal days are exactly the number required to make
up the difference between the normal lunar year of 354 days and the civil year of 365 days. The fifty-nine deities
represent, then, the lunar year rounded out to the number of days of the schematic or averaged lunar year, the
civil year. Could the essential duality of the year be more graphically portrayed?
281. CONCLUSION. The origin of the civil calendar suggested herein meets all the requirements stated
above (264). It develops out of the already existing calendar; it is not tied to Sothis, hence its gradual forward
movement is not immediately perceptible; and it provides an explanation for the second and later lunar calendar.
Moreover, it is now possible to appreciate better the history of the names of the months, which we examined in
226-37. It is understandable why the names of the original lunar months remained the same, why those of the
civil months were firstly the same as those of the lunar months but later tended to become different when the civil
year had moved away from its companion, and why the later lunar year took its month names from the civil months,
since the two were components of the dual year. Exactly when the second lunar year was introduced remains un-
certain, but it was probably not too long after the divergence between the two forms of the year became apparent.
A good guess might be to put it in the neighborhood of 2500 B.C. From that date the Egyptians had three calendar
years, all of which continued in use to the very end of pagan Egypt.
56
oi.uchicago.edu
EXCURSUS A
THE TRANSFER OF FEASTS FROM THE LUNAR TO THE CIVIL CALENDAR
282. In earlier pages there have been scattered references to the transfer of feasts from the lunar to the
civil calendar. We have seen, for example, that there was a fixed wig-feast and a movable wig-feast, and a fixed
feast of Renutet and a movable feast of Renutet. The fixed wig-feast always fell on I _ht 18 civ., while the fixed
feast of Renutet took place on I ,mw 1 civ. The movable feasts, being determined by the original lunar calendar,
in which that of wag fell on I ;ht 13 (185) and that of Renutet on IIII prt 15 (249), might conceivably have fallen
on any day of the civil year. As we shall soon see, there is evidence that the practice of giving an originally
lunar feast a fixed place in the civil year was carried out on a considerable scale, and it is in that phenomenon
that the explanation lies of Gardiner's theory that the months of the civil year once began with Mesore. To that
theory we have briefly referred before. Since it still has vitality, though in a somewhat different form, it will be
useful to review it again here and to propose a solution to the problem it purported to solve more in keeping with
the calendrical theories presented in this work.
283. Like Brugsch, Gardiner believed that the names of the months derived from the important feasts cele-
brated in them. He, however, adduced several examples to show that in earlier times the eponymous feasts of
at least four months were celebrated on the first day of the following month. This suggested to him that at one
time all the months occupied a position one place lower than in late times and that Mesore had once been the
first month of the year. He found confirmation of this in the Ebers calendar, which begins with wp rnpt, another
name for the month of Mesore. Gardiner marshaled the following evidence:
1. In two tombs of the 18th dynasty (Uc-m-ht and Nfr-htp) the feast of Renutet fell on I Emw 1, while
the month-name Pharmuthi is that of IIII prt.
2. In the third year of Ramses X the festival of Epiphi was celebrated on IIII mw 1 (or 2), while Epiphi
is the name of III mw.
3. In the thirteenth year of Ramses IX mswt RC-Hr-hty, the feast of the birth of Re-Horakhti, fell on
I bt 1, while Mesore is the name of IIII mw.
4. At Illahun in the Middle Kingdom and at Medinet Habu under Ramses III the feast of Hathor fell on
IIII 3ht 1, while Athyr is the name of III ?ht.
5. The festival of nhib kw is the same as that of k' hr k3.
2
The nhb kw-feast in both the Middle King-
dom and the New Kingdom was celebrated on I prt 1, while Choiak is the name of IIII ht.
6. The Ebers calendar confirms the theory that once the year began with Mesore.
284. Gardiner could offer no explanation of why the shift forward of the months, or of their eponymous feasts,
should have taken place. Eduard Meyer, who approved of Gardiner's theory, thought he found an explanation.in
the shift forward through the centuries of the summer solstice away from the day of the rising of Sothis. In 4241
B.C. the solstice took place but a few days after July 19, on which Sothis rose. From the solstice came the name
of mswt RC, at this time correctly the first month of the year. By ca. 1300 B.C., however, the solstice had moved
to a point eighteen days before the rising of Sothis, so that at this time the months were moved forward to keep
the solstice in Mesore.
3
285. In his Zeitrechnung Sethe subjected the whole question to a careful analysis and showed the great dif-
ficulties in the way of an acceptance of Gardiner's and Meyer's theories. He proposed instead that no shift had
ever occurred but that a month was named for the feast to which it led up and which actually fell on the first day
of the next month. Thus, for example, the feast of Renutet, which according to Gardiner's theory should have
shifted forward from I Emw 1 to IIII prt 1, continued to be celebrated on I Emw 1 right down to the Roman
period; but this would be quite correct, because the preceding month, which led up to I Emw, was named
Pharmuthi.
5
286. If we re-examine the theories of Gardiner and Meyer, on the one hand, and of Sethe and Weill, on the
other, we soon discover that neither is free from objections. The difficulties that confront Gardiner's six
57
oi.uchicago.edu
points are these:
1. The feast of Renutet never moves to IIII prt 1 as required by theory.
2. It is now certain that the date of the Epiphi feast was IIII .mw 2.6
3. The feast of mswt RC never moves to IIII gmw 1 as required by theory. In the Ptolemaic period it
still fell on I ht 1 (236).
4. In the calendar of Edfu the whole month of III jt (Athyr) is termed "The feast of the [Lady] of
Dendera." In the same calendar a special festival of Hathor begins on III ht 29 and ends on IIII ht 1.
5. It is at best quite doubtful that the nlhb kw festival and that of k hr k) are one and the same.
Gardiner based his conclusion on a scene in the tomb of Paheri at El Kab. One inscription names the scene as
being part of the feast of nhb k~w, while another apparently refers to the feast of k hr kD. But the signs which
Gardiner would read as the latter ( L- o
)
are given by the W6rterbuch as a variant writing of nhb kw, which
at times appeared as J JU. The only date now known for the feast of ki hr k-, with the name certain,
is IIII ht 1 in the calendar of Esna.
1 0
6. The Ebers calendar is most satisfactorily explained as equating the original lunar calendar with the
civil year, and the reason for the appearance of wp rnpt at its head is simply that that event controlled the lunar
year (217-18).
287. If we now turn to the Sethe-Weill theory, we run into comparable difficulties. The only dates ever given
for most of the eponymous feasts fall within the proper months, while only a few fall on the first of the next
month. Weill would argue, for example, that the feast of Thoth began on I ht 19 and ended on II Jht 1; but for
this latter date there is not the slightest evidence in any temple calendar.
1
1
288. The difficulties that confront both theories suggest that a new approach to the original problem is neces-
sary. It is easily possible to dispose of three of Gardiner's six points, but the remaining three still call for ex-
planation. The Ebers calendar cannot be regarded as proof of either a shifting year or a list of feasts falling on
the first day of the month after that to which they give a name. The feast of nlb kw and that of k lr ka are
not demonstrably the same, and the latter falls properly in its month at Esna. The feast of mswt Rc on I ht 1
has already been discussed. It is the companion feast to wp rnpt and originally meant the day of the rising of
Sothis; but when wp rnpt came to mean also the first day of the civil year, so too did mswt Rc (237). There
remain the three feasts of Hathor, Renutet, and Epiphi, which were all apparently celebrated out of their proper
months. How did this come about ?
289. The explanation, I believe, lies in the transfer of feasts from the lunar to the civil calendar. The feast
of Renutet in the lunar calendar was a full-moon feast. At the time when it was given a fixed day in the civil
year we may suppose that IIII prt 15 lunar was the same day as I smw 1 civ. or, as seems somewhat more likely,
was near the latter date and that the first day of the month was adopted as a more significant and appropriate
day. In order for the equation to be exact, the month of thy would have to begin on about I -ht 20 civ. As may be
seen by reference to Figure 20, (p. 54), this could take place when the civil year had moved sufficiently forward
after its introduction. The same explanation can account for a feast of Hathor on IIII tht 1, though here we are
ignorant of what lunar day was involved.
290. The fact of such a transfer is beyond question. We can be certain that there was a feast of Hathor in the
third lunar month of the year and a feast of Renutet in the eighth lunar month if for no other reason than the fact
that each feast named its month. Why the transfer was made is much less certain. Presumably it was the desire
to have a definite and fixed date, known in advance, that could be prepared for with the minimum of last-minute
confusion. It may very well have been that the fixed feasts actually supplanted their lunar prototypes while the
original lunar year and the civil year were still running concurrently and that it was not until the civil year had
moved away from nature and the later lunar calendar had been introduced as its companion that the lunar feasts
of the original lunar calendars were revived. From then on one might have two dates for each festival, one fixed
to the civil year, the other determined by the lunar year, with varying dates in the civil calendar.
291. It will be noted that the feast of Epiphi was omitted from the above discussion. That is because ipip is
the name of III 9mw not in the original lunar year, where ipt ihmt occurs, but in the later lunar calendar. We
have seen that the names of the months of this calendar were the same as those of the civil months (231).
58
oi.uchicago.edu
EXCURSUS A
Presumably, then, there was a lunar feast of ipip on some unknown day of the lunar month. Reference to the
25-year cycle will show that, if the feast began on almost any day after the sixth lunar day, it would have been
possible for it, in some year of the cycle, to have fallen on IIII Emw 2. In other words, Gardiner's second point
is best explained as double-dating in terms of the lunar and civil calendars.
1 2
292. It is not to be supposed that there was a complete and wholesale transfer of feasts from the lunar to the
civil calendar at one given time. No doubt it was a gradual process over the early years of the dual calendar.
It is clear, for example, that the wig-feast was not transferred at the same time as that of Renutet, because the
former took place in some year when I iht 13 lunar was the same day as I iht 18 civ., that is to say when the
month of thy began on I iht 6, while for the latter thy would have to begin around I iht 20. But over the years it
seems likely that there was a more or less complete shift of all the early lunar feasts. It is possible to give one
demonstration of this from a calendar of the temple of Edfu.
1 3
293. At the beginning of the calendar there are long lacunae, so that it seems likely that there are some
feasts missing in the first month. In the following tabulation I list only the days on which there were feasts, with
the number of days' duration of each feast.
1 4
I t (Thoth): 1, 12, 13, 18-20, 21.
II " (Phaophi): ?, 5, 6, 8, 19-Athyr 3, 30-Athyr 5.
III " (Athyr): 1-30, 23, 29-Choiak 1.
IIII " (Choiak): 9-26, 28, 29-30.
I prt (Tybi): 1, ?, 5, 7, 9, 15, 19-21, 25, 28-Mechir 3.
II " (Mechir): 4, 9, 21, 21-30, 28-29.
III " (Phamenoth): 1, 15, 24.
IIII " (Pharmuthi): 1, 4, 2 lunar-21 lunar, 28.
294. It will be observed that the feasts of this calendar occupy by no means all the available days, and at
first glance there would seem to be no order underlying any of them. Closer scrutiny, combined with a lunar
day given in terms of the civil calendar, show otherwise. In the month of IIII prt is the entry * o
On "On the second lunar day of this month was born Horus, son of Isis and Osiris."
1 5
A
- I AJ^ 7
later entry in the calendar under III mw 4 states: "Horus, son of Isis and Osiris, was conceived
He was born on IIII prt 28."
1 6
That IIII prt 28 was indeed the second lunar day is confirmed by the calendar
of Dendera, which has the entry: "IIII prt, pdntyw of this month. At the coming of the third hour, procession
by Hathor, Lady of Dendera, with her Ennead. Resting in the Birth House. Day 28, likewise. Total, two days."
The calendar of Esna hints at the same equation with the entry in IIII prt: "Performing the ritual of the divine
birth of Horus on the second lunar day of this month. IIII prt 28: Feast of Horus, son of Isis."18
295. This was, then, a tradition widespread and of long standing, since it was incorporated in three temple
calendars, all of varying dates of composition. Horus was born on the second day of the lunar month, and at the
time when this feast was fixed in the civil year it fell on IIII prt 28.
296. I cannot resist breaking into the flow of my argument with a brief digression on the connection between
Horus and the moon, as it opens up new paths which invite exploration. Why was Horus born on the second
lunar day? We have seen that on that day the new crescent normally appeared (45). Moreover, the tutelary
deity of the second lunar day is Hr nd it.f, "Horus, avenger of his father."
1 9
That suggests some connection
with Osiris. A Dendera text is so important and illuminating that it must be quoted in as full a form as its state
of preservation permits.
2 0
297. "On this beautiful day the entire land is in joy. All the gods and goddesses rejoice, men are in festival,
the sun-folk are in jubilation, mankind goes forth in gladness in the feast of Choiak, the 24th of the month.
S298. "Horus purifies his father, Osiris, with water; he censes his body with [incense].... He goes forth
from the House of Gold, he being an august spirit ............
299. "They .... the work of Anubis. He receives pure ... from the hands of the sm-priest, divine ... from
the hands of Wadjet, oil from the hand of Anubis. The joy-amulet of gold is placed at his throat. He receives
Adh as the goodly work of Horus, Lord of the Laboratory. His salve is Shesmu's, being pure ... . ........
The 104 amulets of every genuine precious stone are tied for him .........
59
oi.uchicago.edu
300. "..... .. The goddesses adore him on day 24. After he rests in his evening bark, he goes around this
august temple in the ninth hour in the night. He rests in his sarcophagus at the south of this lake. Ceremonies
are done for him .......... The gods are his protection, and all the goddesses are protection around him. He
shines forth from his temple at the setting of the sun on day 25. When he has gone around his city in peace, the
citizens of Dendera are in festival.............
301. " ...... They (the gods) are in his following. Anubis opens for him the way of good ways. He is like
Atum in his setting, when he rests in his beautiful sarcophagus in the necropolis. Hathor, the Mistress of the
West, is his protection. He awakes from sleep, and he flies as the phoenix (bnw), and he makes his place in the
sky as the moon. He receives offerings with Atum, when he enters Dendera in the night of it hr hwt (the fifth
lunar day) to snt (the sixth lunar day) with the goddesses of Upper and Lower Egypt. He finds his sister Isis as
Queen of the Gods of Egypt, there being no command except hers. She places her brother as Ruler, she being
(the female) Ruler, and her son Horus as King of the Gods."
302. Now the calendar of Edfu has the following entry after Choiak 28: p " The feast
of ibt hlr bhwt (the fifth lunar day). Performing his rites.2 If Choiak 28 is the fifth lunar day, then Choiak
24 is psdntyw. If we refer this to the Dendera text, the underlying symbolism is immediately apparent. The
dead Osiris is the vanished moon. His day of embalmment is the day when the moon is never visible. But on
the following day, Choiak 25, he shines forth from his temple at sunset, just as the crescent moon normally ap-
pears on the second day of the lunar month at sunset. If this be the resurrection of Osiris, which it certainly
seems to be, since he wakes from his sleep and takes his place in the sky, then it is small wonder that Horus,
son of Osiris, is born on the same lunar day on which his father is reborn.
303. The better known version of the Osiris myth has, to be sure, his embalmment on the 24th; but his burial
does not take place until the 30th.22 It is not my intention to pursue this fascinating topic any further in these
pages. It is clear, however, that, while the relation between Osiris and the moon has been known for some time,
not all the ramifications of this relationship have as yet been explored. Nor has the connection between Horus
and the moon been suspected.
2 3
But it is time to be done with digression and return to the Edfu calendar.
304. We have established that IIII prt 28 was a second lunar day, which means that the 27th was psdntyw.
Now if we work forward from that date, using the unit of either 29 or 30 days, we soon encounter some astonish-
ing coincidences. There is no feast in III prt on either the 27th or the 28th, but in II prt there is one beginning
on the 28th which lasts two days. I prt has a six-day festival beginning on the 28th. IIII iht has a two-day feast
on the 29th. III 3ht has a three-day feast beginning on the 29th. II ht has a six-day feast beginning on the 30th.
The first month of the year has no significant entry, except that there was a festival of indeterminate length
starting on the first day.
305. Let us now make the assumption that at some unknown time the first day of the lunar year and the first
day of the civil year fell on one and the same day, and that this coincidence was taken as the starting point for
the transference into the civil calendar of certain lunar feasts which began on psdntyw of the following months.
Suppose the following entirely possible sequence of 29- and 30-day months occurred: 30, 29, 29, 30, 29, 30,
29, 30 (the last two might conceivably be 30, 29). We should then expect to find feasts on the following days of
the civil year: I t 1, II .t 1 and 30, III ht 29, IIII _t 29, I prt 28, II prt 28, III prt 27 or 28, IIII prt 27. There
is no evidence in the Edfu calendar for a feast on III prt 27 or 28, and the IIII prt feast actually took place on the
28th, the second lunar day. A lacuna prevents certainty in placing a feast on II t 1, but there was one on some
day before the 5th, and the 1st seems most likely. Granting that, we should then have actual feasts for all the
proposed dates, and important ones too, since the smallest duration, where the information is available, is two
days, and two of them last for six days each.
2 4
To my mind, the conclusion is forced upon us that at one time
the above assumption was fact. That is, the feasts we have traced out in the Edfu calendar have their present
dates because they were originally lunar and were fixed in the civil calendar in the course of one year.
60
oi.uchicago.edu
EXCURSUS B
THE MEANING OF TPY RNPT
306. In the present work it is proposed that tpy rnpt meant originally the first day of the original lunar
year (151). Heretofore the generally accepted interpretation has been that it was the specific name of the first
day of the first month of prt (I prt 1).
1
This was suggested by Brugsch,
2
approved by Sethe,
3
and reaffirmed by
Gardiner.
4
The argument may be summarized briefly as follows: The Sed-festival and the nlhb kiw-festival
were both celebrated on I prt 1. There are texts which apparently date the celebration of both festivals to
tpy rnpt also. Therefore tpy rnpt and I prt 1 are one and the same day.
307. The date of the Sed-festival has been studied by Borchardt. He contended that Ramses II, the only king
whose Sed-festivals are associated with the date I prt 1, used that date to proclaim (sr) to the country the com-
ing festival, which then actually took place on the anniversary of his accession, ten months later, in the follow-
ing regnal year. At the time he wrote his study, Borchardt knew of only two Sed-festivals of Ramses II, the fifth
and sixth, which had been so proclaimed. Evidence for the proclamation of the ninth festival in year 54, I prt no
day, the tenth in year 57, I prt 17, and the eleventh in year 60, I prt 17, is now at hand. These dates head three
separate horizontal single-line inscriptions on the east side of the passage through the first pylon of the temple
of Armant. On the opposite side of the passage there are the remains of three similar lines, mentioning I prt 1
for the years 51, 63, and 65 (or 66), respectively.8 Gardiner has stated, without discussion, that these are the
dates of proclamation of the eighth, twelfth, and thirteenth Sed-festivals of Ramses II;
9
and it must be admitted
that they may be that and nothing else, although little remains of the inscriptions beyond the dates. What does
remain, however, is just enough to cast some suspicion on the proposition. On the first side, where the lines
are preserved in full, following the dates comes the formula etc. On the opposite side, however, the first
sign after the dates is clearly }. In any event, twice the date of proclamation is the 17th day of I prt. This fact
can easily be reconciled with the theory that an announcement was all that was involved, but it is difficult to deal
with on the theory that the Sed-festival began on the 1st.
308. If now, following Borchardt, we assume that the actual celebration began on the anniversary of the king's
accession, and thus on the first day of the following regnal year, we should have the eighth to the thirteenth
festivals celebrated at three-year intervals beginning in year 52. We already know that the first Sed-festival
of Ramses II took place in year 30, the second in year 34, the third in year 37, and the fourth in year 40. The
proclamations of the fifth festival in year 42 and of the sixth in year 45, both on I prt 1, would then mean cele-
brations in years 43 and 46.11 To fill the gap between year 46 and year 52 there would be the celebration of the
seventh festival in year 49, but for this there is as yet no evidence. Festivals occurring with regularity every
three years after the first interval of four years are much more likely than is a two-year interval between the
fourth and the fifth.
309. The available evidence for the celebration of the first Sed-festival of Ramses III clearly favors the
anniversary of the accession day. Gardiner's collation of Pap. Turin 44: 18-19 gives "year 29, 1111 prt 28,"
as the date on which the vizier To sailed north after he had come to take the gods of the South to the festival.
1 2
The arrival of the bark of one of these deities, Nekhbet of El Kab, is commemorated in the tomb of Setau, and
the legend above the scene includes the phrase sp tpy hb sd, "first occurrence of the Sed-festival."13 Now if
the festival were actually to be celebrated on I prt 1, either the southern gods took eight months to reach the
Delta capital or they spent some time there before the festival actually began. On the other hand, with the acces-
sion anniversary of year 30, I Emw 26, as the feast day, the gods would have had nearly a month to voyage to the
North, ample time for the journey, but not so long that they would be absent unduly from their homes.
310. Persuasive likewise for dating the Sed-festival to the anniversary of the accession day is the fact that
the last day of the first festival of Amenhotep III was III Emw 2, exactly eight months after the beginning of
his 30th year on III hjt 2. That the festival lasted exactly eight months is much more probable than that it
lasted six months and two days, its duration had it begun on I prt 1. The third festival of Amenhotep III, in any
61
oi.uchicago.edu
event, began before I prt 1, as we learn from the tomb of UJrw.f.
1 5
The text over the rope in the scene of erect-
ing the dd-pillar reads as follows: "Erecting the d-pillar by the king
himself at daybreak of the Sed-festivals."
1 6
The erection of the dd, according to the Medinet Habu and later
calendars, invariably took place on IIII lht 30; this day must, then, have been included in the period of celebration
of the festival of Amenhotep III. Sethe's contention that in this passage hd ti means "day before"
1 7
cannot be ac-
cepted, since the same expression occurs at Soleb in connection with the illumination of the baldachin on IIII prt
26. Previously in the same inscription we are informed that the ceremony of illumination lasted from IIII prt 26
to I smw 1,18 and r hd t3 n .hbw sd can have reference only to the morning of the 26th, not to the day before.
311. There is, then, no compelling evidence that the celebration of the Sed-festival began on I prt 1, and the
probability is that it began on the anniversary of the accession day, thus differing with each king. What evidence
can be adduced that it began on I tpy rnpt, "the first day of the year"? Brugsch based his conclusion that it
did so on the following text of Seti I:19
"Thou appearest in thy sedan-chair of the Sed-festival like Re on tp(y) rnpt."
This can easily be a mere comparison of two elements
2 0
and no statement of date at all. Moreover, I is
written, not This may be merely a general reference to the "beginning of the year," not to a specific "first
day of the year" (cf. Wb., V, 270 and 280).
312. In a very fragmentary inscription accompanying an equally fragmentary scene of the Sed-festival in the
funerary temple of Pepi II occurs the group I .21 The conclusion that this certainly gives the date of cele-
bration of the Sed-festival of Pepi II may well be doubted; but, granting that it does, it still remains to be deter-
mined what day is here expressed by tpy rnpt.
313. If we now turn to the nhb kw-festival, we find an accumulation of evidence that it was celebrated on
I prt 1.22 But again the evidence that the day of celebration of this festival was called tpy rnpt is quite unsatis-
23
factory. Brugsch quoted two passages from Dendera as proof. In the first the sun-god Harsamtawi is referred
to as 0f I T T A, "resplendent of body at the beginning of the year of Nhb-k (w)"; in the second the
same deity "shines forth from his shrine in his portable bark on his beautiful feast of the begin-
ning of the year, renduring' in his day of coming that he may see b2 N.b-k(w)." In both instances 6 is
written, and in both instances not the feast of nhIb kiw but the serpent Nhb-kiw is mentioned after tp rnpt, and it
is uncertain whether there is any connection between the demon and the feast of the same name.24
314. Now it is true that the festival of nhb kiw was considered a New Year's feast.
2 5
It is also true that
I prt 1 was, according to the Edfu calendars, the day of the accession of Horus the Behdetite and also a feast of
wp rnpt.26 The beginning of such a regnal year might be referred to as tp rnpt,27 and its first day might be
called tpy rnpt.
2 8
What might be true of the accession day of Horus might also be true of the accession day of
any king whose regnal years were counted from his accession day, so that if his Sed-festival were celebrated
on the anniversary of that day it might also be said to be celebrated on tpy rnpt. This might be the explanation
for the appearance of tpy rnpt on the Pepi II relief; but it might just as well be that he celebrated a Sed-festival
on the day properly termed tpy rnpt, "the first day of the lunar year."
315. We have seen how wp rnpt was in origin the term applied to the rising of Sothis, but that in later usage
it came to denote as well the first day of the civil year, the day of accession, and the day of the king's birth.
Tpy rnpt, originally applied only to the first day of the lunar year, may similarly in later times have been
applied to the first day of any year, such as the civil or the regnal year; but for this there is as yet no conclu-
sive evidence. The interpretation that it ever was limited specifically to I prt 1 should be abandoned.
62
oi.uchicago.edu
EXCURSUS C
THE TWELFTH DYNASTY
316. William F. Edgerton has recently formulated a canon for the 12th dynasty, based upon a reaffirmation
of the validity of the date IIII prt 16 for the heliacal rising of Sothis in the seventh year of Sesostris III. His
conclusion is that this year is to be expressed as 1870 B.C. ca. 6 years, this latitude being the result of a pos-
sible range of the arcus visionis (21) from 9.5

to 8.4

and the possibility that the observations might have been
made somewhere between Illahun (lat. 29.2
) and Heliopolis (lat. 30.1). Edgerton was well aware that there are
references to lunar days in the Illahun archives, and concerning these he wrote:
Several Twelfth Dynasty papyri apparently refer to lunar months in connection with regnal years, months, and
days. Future research may perhaps prove that some one of these texts belongs to a particular reign; and if the
reign proves to be that of Sesostris III such evidence in combination with the Sothic date may enable us to equate
the sixteenth day of the eighth month in his seventh year with a particular day in a particular Julian year B.C.
2
It is the object of this excursus to assign such a definite dating to the Sothic date and then by extension to the
whole 12th dynasty.
3
317. All our lunar data from the Middle Kingdom come from the papyri found in the precincts of the mortuary
temple of Sesostris II at Illahun. The papyri are as yet unpublished. Borchardt devoted much time to their study,
and the dates we have are due to his industry; but it is to be regretted that we have as yet no independent check
upon his results. Regarding the lunar dates he has found he said:
Fuir unsere Zwecke hier sind diese Erwahnungen nicht alle zu brauchen, da in vielen Fallen Jahr und Kalender-
tag uiberhaupt nicht genannt, auch oft nicht oder nur unvollstandig erhalten sind. Dazu kommt noch eine andere
Unsicherheit, die zuerst auch bei der Bestimmung des bekannten Hundssternfruhaufgangs Kopfzerbrechen gemacht
hatte: Die Frage nach dem regierenden Konige. Die Hauptmasse der Papyri stammt zwar aus der Zeit Sen-
wosrets III., aus der Senwosrets II. nur wenige, mehr aus der Amenemhets III. und ganz vereinzelte aus noch
spaterer Zeit. Der Fund is nun zwar schon seit 1899 in Museumsverwahrung, aber infolge der Ungunst der
Umstande trotz einiger Ansatze doch noch lange nicht so bearbeitet, dass man wenigstens bei den meisten seiner
Stdcke angeben kSnnte, unter welchem Herrscher sie geschrieben worden sind. Man ist also vorlaufig noch
darauf angewiesen, fiir die Zuteilung der Schriftstiicke in die Zeit bestimmter Konige, die nur sehr selten im
Schriftstick selbst erwahnt sind, zu allerhand, nicht immer sicheren, Aushilfsmitteln, wie Vorkommen bestimmter
Tempelbeamten, Vergleichung der Handschrift usw., seine Zuflucht zu nehmen.
4
There are four certain lunar dates.5 Three are easily detailed:
A. Berl. Mus., Pap. 10090, recto. Borchardt notes (ibid., p. 45, n. 2): "Tagebuch, mittlere dicke
Schrift."
Year 3, III Emw 16 = lunar day 1
B. Berl. Mus., Pap. 10062 A, recto, iii, 6. Borchardt notes (ibid., p. 46, n. 11): "Tempeltagebuch,
grosse Schrift (Zeit Amenemhets III)."
Year 29, I Emw 16 = lunar day 9
C. Berl. Mus., Pap. 10006, recto, col. ii. Borchardt notes (op. cit., p. 45, n. 3): "Tagebuch, grosse
Schrift."
Year 32, III .t 6 = lunar day 1
Year 32, III ht 7 - lunar day 2
318. The last lunar date, D, consists of the first days of the months of a lunar year (186). The papyrus
(Berl. Mus., Pap. 10056, verso) from which the dates are taken is a temple account from Illahun which lists
alternate months of phyle-priests. A transcription was given in 1899 by Borchardt,
6
but unfortunately the papyrus
itself has not yet been published. Through the kindness of Dr. Gardiner, who sent me a photostatic copy of his
photograph of the papyrus, as well as his own transcription of it, it has been possible for me to check all the
dates, which are given in this fashion:
?bd II Emw 26 nfryt r ?bd III Emw 25
?bd IIII gmw 25 nfryt r hit-sp 31 ,bd I tht 198
h t-sp 31 ibd II ?ht 209 nfryt r ?bd III ht 19
tbd IIII t '19 or 1810 nfryt r tbd I prt 18
63
oi.uchicago.edu
tbd II prt 18 nfryt r ;bd III prt 17
3bd IIII prt 17 nfryt r 3bd I sn3w 16
"II mw 26 down to III smw 25
IIII mw 25 down to regnal year 31, I 3ht 19
Regnal year 31, II ht 20 down to III ht 19
IIII ;ht r19 or 18' down to I prt 18
II prt 18 down to III prt 17
IIII prt 17 down to I smw 16."
319. Borchardt, followed by Eduard Meyer,11 believed that the intervening months began on the day after the
second date of each entry. Since he read I ght 20 as explained above, he got a month of 31 days from IIII mw 25
to and including I St 20, and a sequence of 30, 29, 31, 29, 30, 29, 30, 29, 30, 29, 30. Since the 31-day month
fell at the end of the civil year, it seemed an obvious conclusion to Borchardt that the lunar months simply alter-
nated 29 and 30 days, with an extra day added to the last month to correct it.
320. This theory had to meet serious objections. The gravest is that a lunar month is given 31 days. There
is no evidence in any list of lunar festivals or of days of the lunar month for a 31st day. Any scheme which was
based upon the experience of observation would inevitably reject the idea as fantastic. Moreover, as was dis-
closed by G. H. Wheeler, another interpretation of the text was quite possible and led to a much more acceptable
result.13 The date after nfryt r14 was common both to the official finishing his term and to his successor. This
theory gave the acceptable sequence of 29, 30, 30, 30, 29, 30, 29, 30, 29, 30, 29. Furthermore, I believed that I
had found proof of Wheeler's theory in another extract from the archives of the Illahun temple, wherein one phyle
reported as follows:15 "Report of the first phyle of the staff of this temple which is entering the month. What they
have said is: 'All thy affairs are whole and prosperous. We have received all the chattels of the temple, all the
property of the temple being whole and prosperous, from the fourth phyle of the staff of this temple which is with-
drawing from the month. The temple is fortunate in all good.' " This extract seemed to make it perfectly clear
that at Illahun in the Middle Kingdom the day on which a new phyle entered on its work, the first day of a lunar
month, was also the day on which the old phyle rendered an accounting of its stewardship and withdrew.
321. Both of these theories, however, were based upon Borchardt's incorrect reading of I ht 20. Since the
correct reading is I t3ht 19, it has become necessary to review completely the problem of the dates after nfryt r.
It may be presented as follows, with Possibility I giving the intervening months as beginning on the day after the
date following nfryt r and Possibility II giving the intervening months as beginning on the same day as the date
after nfryt r.
Possibility I Possibility II
Day after Same day
Lunar month begins Duration Lunar month begins Duration
II smw 26 II smw 26
30 29
III " 26 III " 25
29 30
IIII 25
111I " 2 5
30 29
I "~1t 20 I ht 19
30 31
II 20 II " 20
30 29
III - 20 III " 19
30 29
I prt 19 I prt 18
29 ~~ 30
II - 18 II " 18
30 29
29 30
30 29
I Emw 17 I Emw 16
$322. We now have confusion compounded. If we begin the lunar months according to Possibility I, we obtain
a proper sequence with no month longer than 30 days. But the evidence of the phyle report cited above seems to
point to Possibility II as the correct interpretation; and this in turn has one month of 31 days, which is against all
64
oi.uchicago.edu
EXCURSUS C
likelihood. It is clear that for chronological purposes it is safe to utilize only the initial dates of each pair as
being certainly psdntyw. After definite results have been established, it might then be possible to make an intel-
ligent choice between the other alternate dates. To summarize, we have these dates of pgdntyw, with B reduced
from lunar day 9 for ease of calculation:
A Year 3, III smw 16
B Year 29 (Amenemhet III), I smw 8
C Year 32, III b.t 6
D Year 30, II smw 26
IIII mw 25
Year 31, II 'bt 20
IIII bt '19 or 18'
II prt 18
IIII prt 17
323. The next step is to check the lunar dates against one another to determine, if possible, whether they
make up one, two, or more groups. The usual way of checking two lunar dates of psdntyw is to total the number
of days between them, divide by the average number of days in a lunar month, 29.53059 (7,9) or for simplicity
29.53, and see if there is any remainder. Using this method, Borchardt had demonstrated that A and C go to-
gether,16 since they are 10,335 days apart and 350 average months total 10,335.7 days, an excess well within
the limits of lunar variability. A much easier method, however, is by the use of the completed 25-year cycle
(Table 5). We have seen that its dates reflect the movement of the moon with an accuracy approaching 75 per
cent (121), with the inaccuracy in no case greater than one day. Thus one has only to fix one date in the cycle
and count years to the other date to check for agreement, and this is easily done, since in the Middle Kingdom
civil years and regnal years coincided.
1 7
III smw 16 of A fits exactly in cycle year 2. If we call that "year 3"
and count to '"year 29" of B, we must equate that with cycle year 3. But in this year a month begins not on
I Smw 8 but rather on I smw 5. Obviously A and B do not fit together. Years 30 and 31 of D would correspond
to cycle years 4 and 5; but every date of D is then off by two days, so that D likewise does not fit with A. Year
32 of C, however, equals cycle year 6; and in that year a month does begin, exactly as required for agreement
with A, on III 3ht 6. A and C do, therefore, fit together, and the use of the cycle gives the result reached above
by computation.
324. Obviously, since A does not fit with B and D, C also cannot fit with them. I smw 8 of B is to be found in
cycle year 25. If that be "year 29," then the following two years of the cycle, 1 and 2, would be "year 30" and
"year 31." The dates of D agree exactly with the cycle for "year 31" and are off but a day for "year 30." Clearly,
B and D are to be paired.
325. We have now divided our lunar dates into two groups, one of which comprises A and C, the other B and
D. The highest date in the first group is year 32, and in the second year 31. Since the papyri come after the
reign of Sesostris II, one group must belong to Sesostris III and the other to Amenemhet III, as the Turin papyrus
gives the former 30 + x and the latter 40 + x years. We have noted that Borchardt, no doubt as a result of his
study of names or script, assigned B to Amenemhet III. With B must go D, leaving A and C for the reign of
18
Sesostris III. Since he reigned between 30 and 39 years, we have next to check our two groups of dates against
each other to see if there is a possible fit which would give Sesostris III a reign within those limits. Placing C,
as before, in cycle year 6, we find that assuming a reign of 36 years for Sesostris III we can get a fit with D in
cycle years 15 and 16.
326. Against this possibility should also be set the other, on the assumption that Borchardt was wrong in
ascribing B to Amenemhet III, that B and D belong to Sesostris III and A and C to Amenemhet III. Should that be
the case, by setting D in cycle years 1 and 2 and assuming a reign of 39 years for Sesostris III we can get a fit
within a day for A in cycle year 13.
327. We have now done all that we can in a deductive approach to our data. There remains the acid test of
actual computation. This will be facilitated if we can fix the 25-year cycle into the 19th century B.C. To do this
I calculated six lunar dates for the year 1847 and six for 1846. Ten of the twelve calculated dates fitted exactly
into cycle years 6 and 7. As a working basis, then, we may set cycle year 1 in 1877, 1852, and 1827 B.C.
328. We may now test both of our possibilities in terms of actual years. We have accepted Edgerton's find-
ing that year 7 of Sesostris III must be 1870 6 years. According to our first hypothesis A and C are dates of
65
oi.uchicago.edu
that king. A fits cycle year 2 and should therefore be 1876. If year 3 is 1876, then year 7 would be 1872, a very
good fit. According to our second hypothesis B and D are dates of Sesostris III. If we place D in cycle years 1
and 2, we have the years 1852-1851 for years 30-31 of Sesostris, which would place his year 7 in 1875, again
within the limits accorded it. Both hypotheses must now be tested by computation. The results are as follows:
Hypothesis I.
Year 3 is 1876
Year 32 is 1847
Year 29 is 1814
Year 30 is 1813
Year 31 is 1812
Hypothesis II.
Year 29 is 1853
Year 30 is 1852
Year 31 is 1851
Year 3 is 1840
Year 32 is 1811
A and C are Sesostris III
B and D are Amenemhet III
Year 7 of Sesostris III is 1872 B.C.
Given Date Calculated Date
III smw 16 = III Emw 16
III t 6 = HII 3t 6
I mw 8 I mw 7
II mw 26 - II mw 26
111 gmw 25 111I gmw 25
II 3_t 20 II iht 19
IIII t 18' = IIII _.t 18
II prt 18 II prt 17
IIII prt 17 - 1111 prt 17
B and D are Sesostris III
A and C are Amenemhet III
Year 7 of Sesostris III is 1875 B.C.
Given Date Calculated Date
I gmw8 - I mw8
II gmw 26 II gmw 27
IIII gmw 25 1IIII mw 26
II !t 20 II 3ht 21
Illl bt "191 IIII ht 20
II prt 18 = II prt 18
IIII prt 17 = IIII prt 17
III gmw 16 III Emw 17
III ?ht 6 III jht 7
329. The probability is all in favor of the first hypothesis being correct. Not only is B a date of Amenemhet
III, as Borchardt had decided, but the percentage of agreement with the astronomically correct dates is much
higher than in the case of the second.
1 9
Moreover, year 7 of Sesostris in 1872 is nearer to the median of
Edgerton's date of 1870 1 6 years. The available lunar data may be said, then, to have fixed the 12th dynasty to
the year down through year 45 of Amenemhet
III.
20
330. I have pursued this comparison of hypotheses to the end, as it has value in itself and as for long it
seemed the only basis for assigning the two groups of lunar dates to the proper kings. Happily, confirmatory
evidence of the greatest importance is now available, since G. Posener has succeeded in reading the name of the
mty n s' in line 4 of Pap. 10056, our lunar date D, as Mkt s Nhti-nb.
2 1
This same phyle-leader is mentioned in
Ku pr 22"
Kahun papyrus IV 1: 5a-6a,22in association with a "year 40" which cannot be other than Amenemhet III's. Thus
the occurence of his name in Pap. 10056 places that too in the same reign.
331. The results obtained from the lunar dates also justify the astronomical correctness of the forecast of
the heliacal rising of Sothis, which must have taken place on July 17, 1872 B.C., the Julian equivalent of IIII prt
16. If we now calculate the arcus visionis (j3) which would give this result, we have the following:
Heliopolis (lat. 30.10) ranges from 9.4

to 8.7
Memphis (lat. 29.90) " " 9.5 to 8.8
Illahun (lat. 29.2
) " 9.7

to 9.0
The experimental data we now have for /3 give it a range from 9.4

to 8.6

(21). The conclusion may thus be
offered that the place of observation in the Middle Kingdom was probably not Illahun but may have been either
Heliopolis or Memphis.
66
A
C
B
D
B
D
A
C
oi.uchicago.edu
EXCURSUS C
332. Before proceeding to an attempt to fix the end of the 12th dynasty and the reign of Amenemhet IV, it will
be interesting to decide, if feasible, between the two possibilities given above (321) for the dates after nfryt r in
D. With years 30-31 of Amenemhet III fixed to 1813-1812 B.C., it is easy to check both possibilities against the
astronomical dates.
Possibility I Possibility II
Given Date Calculated Date Given Date
III smw 26 III smw 25 = III smw 25
I iht 20 I ?bt 19 = I t 19
III ?_t 20 III t 18 III ht 19
I prt 19 I prt 18 = I prt 18
III prt 18 III prt 17 = III prt 17
I smw 17 I smw 16 - I smw 16
333. Here the evidence is overwhelmingly in favor of Possibility II, with five of six given dates agreeing with
calculated ones. Possibility II, we recall, is that toward which the evidence of the phyle report pointed, namely,
that the last day of one phyle was the beginning of the month for the following phyle. But Possibility II also leads
to a month of 31 days, from I it 19 to II bt 20. This points again to Borchardt's theory that the lunar calendar
of the Middle Kingdom used schematic months and that once a year, at the end of the civil year, observation was
used to correct the calendar, with now and again a 31-day month resulting. There are two serious difficulties to
this theory, the first of which is that the lunar calendar concerned is not the later one, tied to the civil year, but
the original one, governed by Sothis (186). Consequently any yearly observational correction would almost cer-
tainly be made with the beginning of the month thy, which at this time would fall around the first month of smw.
The second difficulty is even more serious. The correction resulting in a 31-day month does not agree with the
astronomically correct date. Instead, the extra day results in beginning the next month one day after the morning
on which the crescent was certainly invisible. It can hardly be argued that poor observational conditions might
have caused the extra day. Whether conditions were good or bad, the fact remains that the old crescent could not
have been seen and there could be no excuse for giving a 31st day to the month.
334. A schematic calendar, corrected yearly, is clearly out of the question. Equally clearly, one of the dates
leading to the 31st-day month must be wrong. The first of them, I t 19, agrees exactly with the calculated date.
The second, II 1ht 20, is one day later than the calculated date. With little hesitation I suggest an emendation of
the latter to II ~t 19, which agrees with observation and gives a normal month of 30 days.
335. These last paragraphs do not constitute argument in a circle. Once the 30th and 31st years of Amenemhet
III had been fixed by using all the lunar dates of A, B, and C and the certain ones of D, it became feasible to
examine the remaining dates of D in the light of probability. To my mind it is almost a certainty that Possibility
II is the correct choice, and it is likewise almost a certainty that observation and not a schematic calendar de-
termined the beginnings of the months. That being the case, an emendation to dispose of a 31-day month which
cannot occur with observation seems obligatory. Combining the results of our work on lunar date D leads to the
following:
Calculated Date Given Date Duration
II T mw 26 = II mw 26
29
I ?k.t 19 - I .t 19
II , 19 - II ,, <19>
III , 18 III , 19 30
IIII , 18 = IIII , r181 29
I prt 18 = I prt 18
II - 17 II , 18
III , 17 = III , 17
IIII , 17 - IIII , 17 30
I Smw 16 - I Nmw 16 29
Ten out of twelve given dates thus agree with calculated dates-quite a satisfactory result.
336. We have seen that the wig-feast dated II smw 17 in year 18 of Amenemhet III took place on the thirteenth
67
oi.uchicago.edu
day of the lunar month of thy (182-85). On the assumption that this was the correct lunar day for the feast of
wig, two days before the full-moon feast of thy, just as the fixed feast of wg in the civil year preceded the fixed
feast of thy by two days, it is possible with the help of another dated wag-feast to assign to the year the beginning
of the reign of Amenemhet IV and by extension the end of the 12th dynasty. This important date occurs on the
recto of an Illahun papyrus fragment in the Cairo Museum, No. 58065, and I have to thank the Museum authorities
for permission to publish a photograph of the papyrus on Plate VI B. I give a transcription of the first three lines,
which are all with which we need to concern ourselves, the rest being merely amounts of various offerings. Im-
portant for our purpose is that in "year 9" of some king offerings are to be made to Sesostris II in his mortuary
temple "on the day of the wg-feast, which is to occur on II smw 29.~25
-,- 0
337. The problem, then, is to assign the date II smw 29 to such a year in the 12th dynasty that it is equivalent
to I bt 13 in the original lunar calendar. We recall that the rule regulating the original lunar calendar is that
the lunar month following wp rnpt was intercalary if it began within eleven days of the rising of Sothis. The maxi-
mum number of days that the wg-feast could be distant from the day of rising would be eleven days of the month
wp rnpt, thirty days of the intercalary month, Dhwtyt, and thirteen days of thy, or a total of fifty-four. Fifty-four
days before II smw 29 is I smw 6. The rising of Sothis would then have to fall on this date or later for the wig-
feast to occur on II smw 29. We have seen that in 1872 B.C., the seventh year of Sesostris III, prt Spd took place
on IIII prt 16. There are twenty days from this date to I smw 6. Since Sothis rose one day later every four years,
it follows that it rose on I smw 6 some eighty years after 1872, or 1792 B.C. This year is obviously well past the
ninth year of both Sesostris III and Amenemhet III (1834 B.C.). The only succeeding ruler who had a ninth year
was Amenemhet IV, and by elimination the feast must be his. According to the Turin papyrus, Amenemhet III
ruled more than forty, but less than fifty, years. His year 40 would be 1803 B.C. Let us now, beginning with this
year, construct a table which will take into account the recession of the rising of Sothis by one day every four
years in the civil calendar, and the beginnings of the lunar months immediately subsequent to each rising. Bor-
chardt had calculated that the four years on which Sothis rose on IIII prt 16 were 1875 to 1872 B.C.
2 6
As a work-
ing basis, always subject to check, we may extend these by quadrennia to year 40 of Amenemhet III, which would
begin the four years in which Sothis rose on I smw 4. The 25-year cycle, which we have fixed to the 12th dynasty,
may then be utilized to derive the lunar month beginnings, again with every important date checked by calculation.
There are, of course, only ten possible years for the wg-feast, from 1794 (forty years for Amenemhet III plus
nine for Amenemhet IV) to 1785 (forty-nine years for Amenemhet III and nine for Amenemhet IV) inclusive. With-
in that range (see Figure 21) there is but one possible year for a feast on II smw 29. That is 1790 B.C., and the
fit is exact.
338. With year 9 of Amenemhet IV fixed to 1790, then Amenemhet III had a sole reign of forty-four years.
27 28
Since a year 46, which must be his, is known, his coregency with Amenemhet IV, which is also known, must
have been at least two years. The reigns and coregencies of the 12th dynasty may be summarized as follows:
Sovereigns Sole Reign Coregency
Amenemhet I 20 10
Sesostris I 42 2
Amenemhet II 32 3
Sesostris II 19
Sesostris III 36
Amenemhet III 44 2
Amenemhet IV 9
Sebeknefrure 4
Total 206 17
68
oi.uchicago.edu
EXCURSUS C
Am. III 40
41
42
43
44
Am. IV 1
2,
3
4
5
6
7
8
9
Sebek- 1
nefrure 2+
3+
0c
00
00
0c
00
prt
Spdt
I s mw
)O cooooxxxxxxxxxxxxxxxxxxxxxx:
0 cooooooooooooooooooooooo -- - -
~o OXXXXXXXxxxxxxxXXXXXXXXXXx
oo0oooooooooooooooooo ----
II smw
~xxxxxx------------ w----------
--------------------------- W-----------------cc
---------------------------------- ccccccccccccc
cxxx---------W------------------
- -- - -- - -- -ccccec
0' 000 oooo00xxxxxxxxxxxxxxxx xxxxxxxxxxxx -w----- -
O o ooooooxxxxxxxxxxxxxxxxxxxxxx xxxx ------ w---
.. 0o0oooooooo000000000000000 -w-----------------
o000000000 0000 000000------ - w -------- CCCOcCCOC C
0 000 OOOOXXXXXXXXXXXXXXXXXXXx xxxxxxx-------w----X
00 000 00000000000000000000--------------w-------------------- C c
00 0000 0000 0 oo0oooo ooWxxxxxxxxxxxxx-CCCCCCC
0.). 0000 00000000000000000000 - W -CC
00000000000000000000000 0000------------w-------------
0 00000 0000000000000000000----w------ ---ccc Cc
0' 00000 ooooooxxxxxxxxxxxxxxx: xxxxxxxxxxxxx-------w=-
0 100d000000 000000000000000----w-------- ccc c 8 80dooooooooooooooooooooooo w---------CCCC2c 8
The days of the individual lunar months are shown so:
iptht.... by- ------(the 13th day by w)
wp rnpt 00000 mnbt ccoc
DIhwtyt xxxxx
FIG. 21.-The last years of the 12th dynasty.
339. The, Turin king-list gives a total of 213 years, 1 month, and 19 days for the dynasty, but we do not know
whether the scribe intended to list sole reigns only or to include coregencies. Griffith had long ago remarked
that a simple emendation of 213 to 223 (20 being easily corrupted to 10 in hieratic) would give a figure which
could include all reigns and coregencies.2 With this emendation and a coregency of two years for Amenemhet
III and IV, we can arrive at exactly 223.
340. The results of our inquiry into the lunar chronology of the 12th dynasty are summarized in the following
table, based on Edgerton's. In the chronology of the second millennium B.C. there is no such thing as absolute
certainty, but I submit that there is strong probability that it is correct. For simplicity, the divergency between
the Julian and the Egyptian years, which coincided in 1984, 1983, and 1982 B.C., has been ignored. It should be
noted, however, that every regnal year after the 10th of Amenemhet I actually began in December or November
of the Julian year preceding the one in the table.
TABLE 8
CHRONOLOGY O~F THE 12TH DYNASTY
Amenemhet 1, year 1
Amenemhet I, year 21 Sesostris I, year
Amenemhet I, year 30 Sesostris I, year
Sesostris I, year 43 = Amenemhet II, year
Sesostris I, year 44 Amenemhet II, year
Amenemhet II, year 33= Sesostris II, year
Amenemhet II, year 35= Sesostris II, year
Sesostris II, year 19
Sesostris III, year 1
Amenemhet III, year 1
Amenemhet III, year 45= Amenemhet IV, year
Amenemhet III, year 46 = Amenemhet IV, year
Amenemhet.IV, year 9
Sebeknefrure, year 1
Sebeknefrure, year 4 (last year of dynasty)
1
10
1
2
1
3
1
2
Julian Egyptian
Year B.C. Year of
12th
Dynasty
1991 1
1971 21
1962 30
1929 63
1928 64
1897 95
1895 97
1879 113
1878 114
1872 120
1843 149
1842 150
1798 194
1797 195
1790 202
1789 203
1786 206
69
Cc
1803
02
01
1800
1799
98
97
96
95
94
93
92,
91
1790
89
88
87
oi.uchicago.edu
NOTES TO INTRODUCTION
1. Carl Schoch in Langdon and Fotheringham, The Venus Tablets of Ammizaduga (London, 1928), p. 97.
The minimum figure is possible only under most favorable circumstances.
2. Ibid., (with reference to his article in Biblica [Rome], January, 1928).
3. Ibid.
4. Cf. Parker and Dubberstein, Babylonian Chronology 626 B.C.-A.D. 45 (Chicago,. 1942), p. 23, and Johann
Schaumberger, Sternkunde und Sterndienst in Babel, Erganzungsheft 3 (Munster, 1935), p. 255.
5. Cf. W. F. Edgerton's remarks in "On the Chronology of the Early Eighteenth Dynasty (Amenhotep I to
Thutmose III)," AJSL, LIII (1937), 192-93.
6. The latest treatment of the problem is that of Walter F. Snyder, "When Was the Alexandrian Calendar
Established?" American Journal of Philology, LXIV (1943), 385-98. He argues plausibly for 30 B.C.
7. The Alexandrian and Julian years are both of the same length, 365 1/4 days, but they vary in the lengths
of individual months and in the location of the extra day in leap year.
NOTES TO CHAPTER I
1. Martin P. Nilsson, Primitive Time-Reckoning (Lund, 1920), p. 150.
2. Cf. F. K. Ginzel, Handbuch der mathematischen und technischen Chronologie, I (Leipzig, 1906), 35; Hutton
Webster, Rest Days (New York, 1916), pp. 178 ff,; and Nilsson, op. cit., pp. 147 ff.
3. K. Sethe, "Die Zeitrechnung der alten Aegypter im Verhaltnis zu der der anderen Volker," Nachrichten
von der Koiniglichen Gesellschaft der Wissenschaften zu Gottingen, Philologisch-historische Klasse, 1919-20, p. 289.
I shall refer to this work hereafter as "Sethe, Zeitrechnung," with the pagination of the Nachrichten, which has
pp. 287-320 in 1919 and pp. 28-55 and 97-141 in 1920.
4. Nilsson, op. cit., p. 169.
5. Richard Lepsius, Die Chronologie der Aegypter. Einleitung und erster Theil: Kritik der Quellen (Berlin,
1849), p. 157. That Lepsius here meant "new crescent" and not " conjunction" is clear from his use of the same
term on p. 158, n. 1, where he speaks of the beginning of the Greek calendar, which certainly started with cres-
cent visibility.
6. Ludwig Ideler, Handbuch der mathematischen und technischen Chronologie, I (Berlin, 1825), pp. 93-194.
7. H. Brugsch, Materiaux pour servir a la reconstruction du calendrier des anciens egyptiens (Leipzig,
1864), pp. 58-60.
8. E. Mahler, "Das mittlere Reich der agyptischen Geschichte," ZAS, XL (1902), 79, and again in "Sothis
und Monddaten der alten Aegypter," Actes du XIVe Congres international des orientalistes, II, No. 4 (Paris,
1907), pp. 39-41.
9. K. Sethe, "Die Spriche fuir das Kennen der Seelen der heiligen Orte," ZAS, LVII (1902), 35, 48, and more
recently in Ubersetzung und Kommentar zu den altagyptischen Pyramidentexten, IV (Gliickstadt, 1935), 16.
10. Zeitrechnung, p. 289.
11. First reference cited in n. 8.
12. D. R. Fotheringham, "Some Considerations regarding Professor Petrie's Egyptian Chronology,"' PSBA,
XVIII (1896), 101.
13. C. F. Lehmann(-Haupt), Zwei Hauptprobleme der altorientalischen Chronologie und ihre Losung (Leipzig,
1898), pp. 150-51.
14. Eduard Meyer, Aegyptische Chronologie (Berlin, 1904), p. 49.
15. William F. Edgerton, "On the Chronology of the Early Eighteenth Dynasty (Amenhotep I to Thutmose III),"
AJSL, LIII (1937), 195.
16. Lynn H. Wood, "The Kahun Papyrus and the Date of the Twelfth Dynasty." BASOR, No. 99 (October, 1945),
pp. 6-9.
17. Duncan Macnaughton, A Scheme of Egyptian Chronology (London, 1932), pp. 145 ff.
18. OLZ, XXVIII (1925), 620. In n. 2 Borchardt says that this proposition was given orally in April, 1923, to
70
oi.uchicago.edu
NOTES
a session of German Orientalists in Berlin.
19. Hereafter referred to as "Mittel."
20. Mittel, p. 7.
21. Ibid., p. 36; see also pp. 7 and 25.
22. Ludwig Borchardt, "Drei neue Beispiele von Mondmonatsnamen aus der Zeit der 20. Dyn.," ZAS, LXXIII
(1937), 67.
23. H. H. Nelson et al., Medinet Habu III: The Calendar, the "Slaughterhouse," and Minor Records of Ramses
III (Chicago, 1934), P1. 148, 11. 294, 306, 318, 367, 379, 391, and P1. 150, 11. 440 and 452.
24. Mittel, p. 37 and n. 2.
25. Ibid., p. 30, n. 1.
26. Ideler, op. cit., pp. 101-2; Lepsius, op. cit., pp. 130-31; Sethe, Zeitrechnung, pp. 130-38.
27. Sethe, Zeitrechnung, pp. 119-30. For evidence that the Babylonians never began their day in the morning,
as stated by B. Meissner, Babylonien und Assyrien, II (Heidelberg, 1925), 394, and Sethe, Zeitrechnung, p. 121,
with reference to Journ. asiat., 1909, p. 341, n. 4, see now S. Langdon, Revue d'Assyriologie, XXVIII (1931),
14-16, accepted by N. Schneider, Die Zeitbestimmung der Wirtschaftsurkunden von Ur III (" Analecta Orientalia,"
No. 13 [Rome, 1936]), p. 117.
28. Zeitrechnung, p. 130.
29. Ibid., pp. 300-302.
30. I know of only one clear example in which a lunar name has been applied to a civil day. In a building
inscription at Edfu the 18th day of IIII mw is given as ich, the name borne by the 18th lunar day; cf. ZAS, X
(1872), 14.
31. Heinrich Brugsch, Thesaurus Inscriptionum Aegyptiacarum, Vols. I-VI (Leipzig, 1883-91), referred to
hereafter as "Thesaurus."
32. Published in Brugsch, Materiaux . . .calendrier . . . , p. 59, and Recueil de monuments egyptiens, I
(Leipzig, 1862), P1. XXXVIII, 2.
33. Published in Adriaan de Buck, The Egyptian Coffin Texts, II (Chicago, 1938), 322-24. I follow the text
of S 2 C. Cf. the parallel text in Sethe, "Die Spriiche fuir das Kennen der Seelen der heiligen Orte," ZAS, LVII
(1922), 10, 1. 25, and Sethe's remarks on pp. 35 and 48.
34. Thesaurus, pp. 34 ff.
35. Rochemonteix and Chassinat, Le Temple d'Edfou, I (Paris, 1897), 375. Future references will be given
as "Edfou."
36. So Helmuth Jacobsohn, Die dogmatische Stellung des Konigs in der Theologie der alten Agypter (Gliickstadt,
1939), p. 23.
37. Mittel, p. 45, n. 1.
38. Kurt Sethe, Von Zahlen und Zahlworten bei den alten Agyptern (Strassburg, 1916), p. 20.
39. Wb., II, 198, calls this nt a "Bildungselement." According to Sethe, Kommentar. ... Pyramidentexten, III,
318, and IV, 125, it is an ordinal ending after the numeral. (In no writing of either the 6th or the 15th day how-
ever, does it ever have the form .) Already in the Pyramid Texts the name of the numeral six, srs or sis,
had been reduced to a simple _ (cf. Coptic Coo , Co), so that the addition to it of nt permitted a writing like the
following which uses the biliteral sign sn, :of (Pyr. T. 716 a). The earliest occurrence of the numeral as
the name of the 15th day seems to be in the 18th dynasty tomb of Chaemhet (No. 57 at Thebes), cited in the
Belegstellen to Wb., II, 198, 2, where it appears as nl" . It is doubtful that this is still to be read imdt; one
might suspect something rather like rmd-dfw-nt'
.
It is likely that this writing is responsible for the appearance
of nt in the variant #~zv , also occurring first in the 18th dynasty (Sethe, Urk., IV, 112). See my earlier re-
marks under imdt, 36.
40. J. Garstang, Burial Customs of the Ancient Egyptians (London, 1907), P1. IX, and Louis Speleers, Recueil
des inscriptions
4
gyptiennes des Musies royaux du cinquantenaire 1 Bruxelles (Bruxelles, 1923), pp. 22-23.
41. Borchardt, Mittel, p. 20 and n. 1, refers to three unpublished examples from the Illahun papyri in Berlin.
42. Arguably an m-formation from the root spr, "arrive, reach"; cf. Wb., II, 144, 6, "sanctuary, asylum (?),"
71
oi.uchicago.edu
NOTES
and 7, "harbor." On the coffin of Ma the writing is o. Borchardt, Mittel, p. 37, n. 2, sees mspr as an m-forma-
tion based on spr, "rib," and compares the new crescent to a rib, first visible on the third day of the month. As
we see above, however, this is true in only three months out of ten, and further he ignores the importance
attached to 3bd by the Egyptians themselves.
43. If conjunction occurs close to midnight of the last day of the month, it is theoretically possible, given the
exceptionally favorable conditions which result in the minimum hours for visibility, that the new crescent may
be seen at sunset on pdntyw. In all my calculations, however, I have not yet encountered an instance of this.
44. I do not know the maximum time required for crescent visibility in Egypt. It should be less than the 42
hours necessary at Babylon (11), since Egypt is farther south (16). Should it be more than 30 hours, there is
the rare possibility of new crescent being delayed to the fourth day. Again, I have not yet met with such an in-
stance.
45. That is not to say that full moon may occur with equal likelihood anywhere in this range. Only rarely
will it take place at either of the extremes. In the great majority of cases astronomical full moon will happen
either in the night of smdt or the daytime of mspr sn-nw.
46. Mahler, op. cit., pp. 39-41.
47. 0. Neugebauer and A. Volten, "Untersuchungen zur antiken Astronomie IV. Ein demotischer astrono-
mischer Papyrus (Pap. Carlsberg 9)," Quellen und Studien zur Geschichte der Mathematik, Abt. B, Band 4
(Berlin, 1938), pp. 383-406.
48. Since the civil year has 365 days, while 12 lunar months total on the average but 354 days, it is neces-
sary that every second or third lunar year have 13 months, so that the civil and lunar years will remain in
general concurrent.
49. This, as was recognized by the editors of the papyrus, is the simple explanation of the feasts of the great
and the small years in the Beni Hasan calendar. Cf. Percy E. Newberry, Beni Hasan, Vol. I (London, 1893),
Pls. XXIV-XXV and pp. 54 and 61.
50. How the intermediate dates were determined will be discussed in the following chapter.
51. Eduard Mahler, Chronologische Vergleichungs-Tabellen, I (Wien, 1889), 15-38, includes tables for the
conversion of Egyptian dates into the Julian calendar from 747 B.C. to A.D. 451.
52. Neugebauer-Volten, op. cit., p. 402.
53. Memphis has been arbitrarily chosen for calculation purposes.
54. Ptolemy (ca. A.D. 100-178) was well aware of the fact that the moon nearly repeats itself in 25 civil
years. His lunar tables (Mathematike Syntaxis [Almagest] VI.3 [p. 466, ed. Heiberg = tr. Manitius, I, 343]) begin
in 747 B.C. with conjunction on Thoth
2 4
d 44' 17". By 346 B.C. the beginning of his 25-year cycle of conjunction
has dropped to Thoth
2 3
d 59' 44" and 550 years later to Thoth
2 2
d 58' 28". (The symbols ' and " indicate 60ths
and 3600ths of a day.)
55. Op. cit., p. 383.
56. On the whole question of the antiquity of Egyptian astronomy cf. Neugebauer, "Egyptian Planetary Texts,"
Transactions of the American Philosophical Society, XXXII, Part II (New ser., January, 1942), 235-43. His con-
clusion regarding the texts published therein is that their date of origin is in the first period of the Hellenistic
age.
57. Gardiner, Thompson, and Milne, Theban Ostraca (London, 1913), p. 51 (D 31). I owe the correct reading
of the year number to Dr. G. R. Hughes.
58. Thompson had no doubt forgotten that Commodus reckoned his years from the accession of his father,
March 7, 161. Cf. H. Gauthier, Le Livre des rois d'Egypte (Le Caire, 1907-17), V, 166, n. 2.
59. W. Spiegelberg, "Eine neue Bauinschrift des Parthenios," ZAS, LXVI (1930), 42-43. The transliteration
of the document is correct as far as the date is concerned, but in his translation and commentary Spiegelberg
has unwittingly changed IIII prt to III prt. Cf. also Borchardt, Mittel, p. 39.
60. Papyrus Rhind I, i, 10-11. Georg Moller, Die beiden Totenpapyrus Rhind (Leipzig, 1913), p. 14. Cf.
Borchardt, Mittel, p. 23, n. 2, and pp. 39-40.
61. Op. cit., p. 75, n. 14.
72
oi.uchicago.edu
NOTES
62. So in the Medinet Habu Calendar for the feast of Opet, P1. 156, 11. 837 ff.
63. The 27th day of the lunar month has as variant to its usual name the less common one of twn bwy, Wb., V.
359, 12.
64. Mittel, p. 23, n. 2.
65. Mond and Myers, The Bucheum, III (London, 1934) P1. XLIII, 13, Gl; Borchardt, Mittel, p. 40.
66. Brugsch, Thesaurus, p. 924; Gauthier, op. cit., IV, 411; Borchardt, Mittel, p. 40.
67. Edfou, VII, 8; Brugsch, Thesaurus, pp. 256 f.:59 f.; Borchardt, Mittel, pp. 23 and 42.
68. Edfou, VII, 7;IV, 8-9;IV, 2; Brugsch, Thesaurus, pp. 255:46, 266:16, and 271 VI; Borchardt, Mittel, pp. 23
and 42.
69. W. Speigelberg, Die demotischen Denkmller, II: Die demotischen Papyrus, Vol. I, Text; Vol. II, Plates
("Catalogue general . . . du Caire," Vols. XXXIX-XL [Strassburg, 1908], pp. 172-74, Pl. LXIV).
70. The space seems too small for both an amount of grain and a name.
71. Emend to t?
72. There had been a fifth phyle in the Old Kingdom, which must later have been discontinued; cf. Sethe, ZAS,
LIV (1918), 3, n. 5, and Duell et al., The Mastaba of Mereruka (Chicago, 1938), Vol. II, P1. 199.
73. Edfou, VII, 6;IV, 7;II, 26;II, 27;I, 327; Brugsch, Thesaurus, pp. 254: 33,266: 6,269 III, 270 IV, and 271 V;
Borchardt, Mittel, p. 41.
74. Edfou, VII, 5 and IV, 7; Brugsch, Thesaurus, pp. 254: 27 and 266 : 1; Borchardt, Mittel, p. 41.
75. The first part of the temple was completed on a 6nt (lunar date 9); it was finished as a whole on a second
dnit (lunar date 7); new construction was begun on a snt (lunar date 6).
76. In the following chapter we shall discuss whether this cycle represents something new or is a correction
of an already existing cycle.
NOTES TO CHAPTER II
1. On the use of these names for the lunar months cf. 231.
2. "Mesore as First Month of the Egyptian Year," ZAS, XLIII (1906), 136-44.
3. Cf., e.g., Mittel, pp. 50 ff.
4. Illahun Papyrus, Berl. Mus., P. 10056, verso, ZAS, XXXVII (1899), 92 ff.; Mittel, pp. 7, 29 ff.
5. "The Chronology of the Twelfth Dynasty," JEA, IX (1923), 199, quoted and approved by Edgerton, JNES, I
(1942), 310.
6. Loc. cit.
7. Urk., IV, 657.
8. Faulkner, in "The Battle of Megiddo," JEA, XXVIII (1942), 4, 11, argues with force and logic that the day
should be emended to 20.
9. Faulkner renders, "the exact day of the festival of the New Moon."
NOTES TO CHAPTER III
1. An excellent study is the chapter, "Lunar Calendars and the Week," in Hutton Webster, Rest Days (New
York, 1916), pp. 173 ff.
2. J. C. Gatterer, "De Theogonia Aegyptiorum," Commentationes Societatis Regiae Scientiarum Gottingensis
Historicae et Philologicae Classis, VII (Gottingae, 1786), 34, 49, 52.
3. Handbuch der mathematischen und technischen Chronologie, I (Berlin, 1825), 174-77, 190.
4. These two events coincided almost exactly in 3000 B.C.
5. Chronologie, pp. 149-58, 220.
6. Thomas Henri Martin, "Mimoire sur le rapport des lunaisons avec le calendrier des Egyptiens sur la
periode d'Apis et sur la periode de 36,525 ans," Academie des inscriptions et belles-lettres: Memoires
presentes par divers savants (Paris, 1860-64), serie 1, tome 6, partie 2, pp. 441-72.
7. Edward Hincks, "On the Various Years and Months in Use among the Egyptians," Transactions of the
Royal Irish Academy, Polite Literature, XXIV (1865), 59-60.
73
oi.uchicago.edu
NOTES
8. Die Aegyptologie, p. 353.
9. Thesaurus, pp. 234, 476. Brugsch (op. cit., p. 245) seems to suggest a lunar year of twelve months only.
10. Thesaurus, p. 245 et passim; Die Aegyptologie, p. 353.
11. Thesaurus, p. 250 et passim. Brugsch was well aware that the civil year continued uninterruptedly after
the attempted reform by Ptolemy III Euergetes I in 238 B.C., but he thought there was evidence that the reform
itself continued as a separate year. This theory has met with no acceptance, and I leave it as not worth the space
to refute in detail (cf. Meyer, Chronologie, p. 31, n. 1).
12. Chronologie, pp. 4 ff.
13. Ibid., pp. 31 ff.
14. Zeitrechnung, pp. 300-302.
15. Ibid., pp. 302 ff.
16. Ibid., pp. 311 ff.
17. Mittel, pp. 5 ff., 24 (cf. also OLZ, XXVIII [1925], 620, and ZAS, LXX [1934], 98-99). In the very same con-
text Borchardt qualified his clear statement that the original lunar year began with the first month after the rising
of Sothis by placing its beginning "around" the rising of Sothis, or "around" the longest day of the year.
18. Primitive Time-Reckoning (Lund, 1920), pp. 247-48.
19. As primary source cf. E. Pechuel-Loesche, Die Loango-Expedition, III:2 (Stuttgart, 1907), p. 138.
20. On this see H. Frankfort, "Modern Survivors from Punt," Studies Presented to F. L1. Griffith (Oxford,
1932), pp. 445-53; C. G. Seligman, "Egyptian Influence in Negro Africa," Studies Presented to F. Ll. Griffith,
pp. 457-62; C. G. Seligman, Egypt and Negro Africa (London, 1934).
21. Sir W. Willcocks, The Nile in 1904 (London, 1904), Appendix K, Table 41, as quoted in Otto Neugebauer's
article, "Die Bedeutungslosigkeit der 'Sothisperiode' fur die alteste agyptische Chronologie," Acta Orientalia,
XVII (1939), 185, Fig. 2, and 190, Fig. 4.
22. For Memphis the rising of Sothis is about five days later and the beginning of the inundation about ten days
later than at Assuan. Where in the valley the calendar first developed, it is obviously impossible to say.
23. Sethe in his Kommentar to this passage prefers the translation "who renews herself," but I follow Junker
in Giza, III (Wien, 1938), 111-13, and Giza, IV (Wien, 1940), 27, where he demonstrates the meaning of rnpwt,
"year-offerings."
24. Brugsch, Reise nach der Grossen Oase el Khargeh (Leipzig, 1878), P1. XVI, 29 and 33-34 = Thesaurus,
pp. 510-11.
25. Brugsch, Thesaurus, p. 390.
26. Most recently from text 1 by Junker, Giza, IV, 27.
27. "Ein neues Sothis-Datum," ZAS, VIII (1870), 110.
28. F. Ll. Griffith, The Inscriptions of Siat and Der Rifeh (London, 1889), Tomb I, 1. 278.
29. William F. Edgerton, "Chronology of the Twelfth Dynasty," JNES, I (1942), 307-14, is the latest discussion.
30. The Semneh inscription must then have been drafted sometime prior to the 30th year of Sesostris III.
31. This is also the conclusion of Brugsch, Thesaurus, p. 234. Earlier, in Materiaux . . . calendrier .. .,
p. 29, he had come to the opinion that wp rnpt = prt Spdt on the basis that the two were never listed together in
festival lists.
32. Cf. also Sethe, Beitrage zur Altesten Geschichte Agyptens (Leipzig, 1905), p. 63; Zeitrechnung, p. 294; and
especially, H. E. Winlock, "The Origin of the Ancient Egyptian Calendar," Proceedings of the American Philo-
sophical Society, LXXXIII (1940), 457 and n. 31.
33. Pyramid Texts 632 shows clearly that Sothis was identified with Isis early in Egyptian history, and the
association of Isis with the cow (through identification with Hathor) is well known.
34. In the stelas published by Dows Dunham in Naga-ed-D~r Stelae of the First Intermediate Period (London,
1937) there are only three that list feasts: 37, 85, 87. Wp rnpt is included on 37 and 87. Prt Spdt does not occur.
35. E. g., in Lange and Schafer, Grab- und Denksteine des Mittleren Reichs im Museum von Kairo, Vol. I
(Berlin, 1902), there are seven lists of feasts. Wp rnpt is listed on 20005 only, while prt Spdt is found on 20326,
20338, and 20390.
74
oi.uchicago.edu
NOTES
36. E.g., Urk., IV, 538; cf. also 823-24.
37. In col. i, 1. 2, of the recto of the new Cairo calendar of lucky and unlucky days, Pap. 86637, of Ramesside
date, occurs the significant entry 3bd 1 bt i w 1 wp rnpt sn-nw, "I bt 1, the second wp rnpt." In the light of our
theory the situation is crystal-clear. The rising of Sothis would be *wp rnpt tpy, and the application of wp rnpt
to the first day of the civil year would be secondary and correctly designated 6n-nw. I am greatly indebted to
Dr. Abd el-Mohsen Bakir of the Egyptian Museum, who will publish the papyrus, both for calling the passage to
my attention and for permitting me to refer to it.
38. Cf. Junker, Giza, II, 59 ff.
39. Brugsch, Thesaurus, pp. 362 ff., or Drei Fest-Kalender (Leipzig, 1877) may be conveniently referred to.
40. This is, no doubt, a , "the very, very great feast," of Edfou, V, 351:11. Cf. hb wr in 4th dynasty
again, Brugsch, Thesaurus, p. 235.
41. Cf. Borchardt, Mittel, p. 56.
42. Newberry, Beni Hasan, I, 53, and P1. XXIV.
43. We shall discuss later the transfer of feasts from the lunar to the civil calendar. Such a transfer would
still keep the feasts in the same general order of succession.
44. Mittel, p. 34 and nn. 3 and 5.
45. Berl. Mus., Pap. 10016, published by Scharff in ZAS, LIX (1924), 24 ff. He assigned it with some reserva-
tion to Sesostris III, his reason being that the imy-r ~ tp-ntr Sobeknakh't is mentioned in the letter and an official
of the same name and title was the recipient of another letter, apparently to be dated to the 18th year of Sesos-
tris III (op. cit., p. 23 and n.). But neither the name nor the title was uncommon. Moller, Hieratische Palaogra-
phie, Vol. I (Leipzig, 1909), P1. V, assigns the papyrus to Amenemhet III.
46. The dates for year 18 of each king are based on the results attained in Excursus C, .v
47. Mittel, p. 35.
48. Ibid., p. 34, n. 3.
49. According to Brugsch, Thesaurus, p. 393, it had a duration of fifteen days (the last half of the month); but
no source is given. Borchardt (Mittel, p. 34) quotes Brugsch and adds "in Edfu." According to the Dendera
calendar (Thesaurus, p. 365: 10), the ty-feast lasted five days; but Brugsch (op. cit., p. 286) states that it began
there on Thoth 20 and ended on Phaophi 5. Ty was the eponymous feast of the first month of the lunar calendar,
and at least one other such feast, that of Renutet, was celebrated on the day of full moon ($249). Evidence on the
other eponymous feasts is not forthcoming.
50. "Ein neues Sothis-Datum," ZAS, VIII (1870), 108-11.
51. Raymond Weill, Bases, methodes et r6sultats de la chronologie egyptienne (Paris, 1926), pp. 113 ff.; Sethe,
Zeitrechnung, p. 314; and others.
52. "On the Chronology of the Early Eighteenth Dynasty (Amenhotep I to Thutmose III )," AJSL, LIII (1937),
188-97.
53. Borchardt offered only three unpublished cases from the Illahun papyri where pd written with the figure
9 is followed twice by the feast sign and n and once by a lacuna (Mittel, p. 20, n. 1).
54. Borchardt was no doubt misled by an old error of Brugsch; cf. Drei Fest-Kalender, p. viii, n. 1.
55. Recognized long ago by Lepsius, "Geschichtlichkeit der 5ltesten Nachrichten," ZAS, XIII (1875), 150.
56. Col. xviii, 2, Emw; lxi, 4-5, II-1IIII p; lxi, 15, I - II prt; lxxxvi, 8, Emw.
57. Lepsius, op. cit., p. 154.
58. Eduard Meyer, who rejected the fixed year, had to confess that the Ebers calendar was a riddle to him
(Chronologie, p. 48).
59. On the basis of the Turin king list and a reconstruction of the royal annals (Palermo stone, etc.) I would
date Menes to ca. 3110 B.C. I hope later to publish my results, but much remains to be done.
60. Thesaurus, p. 247.
61. "The Origin of the Ancient Egyptian Calendar," Proceedings of the American Philosophical Society,
LXXXIII (1940), pp. 447-64.
62. Ibid., p. 451.
75
oi.uchicago.edu
NOTES
63. Loc. cit.
64. Sethe, Zeitrechnung, p. 311; Weill, Bases, pp. 129-30.
65. Otto Neugebauer, "Die Bedeutungslosigkeit der 'Sothisperiode' fur die alteste gyptische Chronologie,"
Acta Orientalia, XVII (1938), 185.
66. Sethe, loc. cit. and n. 1; Weill, loc cit.
67. Brugsch, Thesaurus, p. 382: 14. According to the Edfu calendar (Edfou, V, 351, 10), there was a feast to
Renutet celebrated there on I prt 7. For a discussion of this see below (250).
68. Sethe, op. cit, p. 312; Edfou, V, 353, 15-16. This may have been in reality the feast of Min; cf. Drioton,
"Les fetes de Bouto," Bulletin de l'Institut d'Egypte, XXV (1943), 7.
69. Sethe, loc. cit.
70. Sethe, op. cit., p. 313; Weill, op. cit., p. 128.
71. Sethe, loc. cit.
72. Based on the figure in Illustrated London News, May 10, 1930, p. 846.
73. Neither is it the sole example of an astronomical year beginning in the first part of October and based on
the culmination of Sothis (Borchardt, "Altagyptische Zeitmessung," in Ernst von Bassermann-Jordan, Die
Geschichte der Zeitmessung und der Uhren, I, B [Berlin, 1920], p. 21). As Borchardt stated, the scales of the
water clock do not fit the reign of Amenhotep III (ca. 1400 B.C.) but reflect the calendarial situation of the civil
year from ca. 1630 to ca. 1510 B.C. Now it was just around the middle of the sixteenth century B.C., in the reign
of Amenhotep I, that the astronomer Amenemhet first invented the water clock (op. cit., pp. 60-63). His ratio be-
tween the longest scale and the shortest scale, 14 fingers to 12 fingers, fits the Karnak clock exactly. It seems
to me quite safe to conclude that the scale on the Karnak clock, fitting as it does the period of the inventor, is
simply another manifestation of Egyptian conservatism, similar, for example, to the decanal calendar ceilings
of Seti I and Ramses IV, which are basically the same though separated by about a century and a half. The cor-
rect months for the time of Amenhotep III may have been painted on the rim of the clock, or a conversion calen-
dar may have been employed.
74. Weill, op. cit., pp. 138-54.
75. De Iside et Osiride 13 c, 42 a.
76. Ibid. 39 bc.
77. "Les fetes d'Osiris au mois de Khoiak," Recueil de Travaux, III (1882), 43-57; IV (1883), 21-33; V (1884),
85-103.
78. ZAS, XLIII (1906), 142.
79. ZAS, VIII (1870), 109.
80. Mittel, p. 22. The year concerned in the text is not the lunar but the civil. Since, however, as we shall
discuss in detail later (230), the names of the lunar months were transferred to the months of the civil year,
without shift in relative position, it is simpler on this page to understand "year" as applying either to the lunar
or civil calendar or to both.
81. Griffith-Petrie, Two Hieroglyphic Papyri from Tanis (London, 1889), P1. IX, 2.
82. Pierre Montet, "Inscriptions de basse dpoque trouvees a Tanis," Kemi, VIII (1946), 35-39 and Pls. I and
II. Of the month fields are preserved only those of III Emw, with the figure of Ipet, the hippopotamus goddess,
and of IIII Emw, with the figure of Re-Harakhti as a falcon and the feast-name, wp rnpt.
83. Cf. Brugsch in ZAS, X (1872), 15. Borchardt quoted this passage (Mittel, p. 23) but failed to connect it
with the others.
84. This meets one of Gardiner's main arguments for his theory that Mesore was once the first month of the
year (ZAS, XLIII [1906], 141). Elsewhere (Excursus A) we shall discuss his theory in more detail.
85. Edgerton (AJSL, LIII, 193) gives the probable years for the rising of Sothis on III Emw 9 in the ninth year
of Amenhotep I as 1544 to 1537 B.C. In 1542, well within this range, a lunar month began, by calculation, one
day before III Emw 9. Since the difference is but one day, it is possible to suppose an error in observation, and
the interpretation given above need not be summarily rejected. Of possibly greater import, however, might be
the fact that this settlement of the Ebers date does not accord with Borchardt's theory of the accession of the
76
oi.uchicago.edu
NOTES
pharaoh at full moon. To be sure the accession date of Amenhotep I is not known for certain (Borchardt, Mittel,
p. 28; Gardiner, JEA, XXXI, 25); but of all the dates offered not one can be brought close to a full moon, if the
ninth year be 1542. At a future time I propose to examine critically this theory, which has been the object of
suspicion by some Egyptologists but has been stoutly defended by Borchardt.
86. Cf. the excellent discussion of the use of the schematic lunar calendar in Babylonia by O. Neugebauer,
"The Origin of the Egyptian Calendar," JNES, I (1942), 400 ff.
87. H. E. Winlock, Excavations at Deir el Bahri 1911-1931 (New York, 1942), pp. 137-40 and Pls. 60-67.
88. There are minor variations in orthography. The third month has the phrase pt h1nc 6bt.s, "heaven with
its stars," before the name of Ht-hr; and the tenth month has prty, the meaning of which is not clear to me,
after Int-hty.
89. Cf. Brugsch, Thesaurus, p. 53, where he discusses the Ramesseum ceiling, to which Senmut's is ances-
tral.
90. Thesaurus, pp. 471-73; Die Aegyptologie (Leipzig, 1891), pp. 359-64.
91. The decanal deities here depicted are a selection of thirty-six out of the fifty-nine divinities of the dual
year, and consequently have a lunar aspect; cf. 273-75.
92. The main differences are in the fourth to the seventh month and in the eleventh month, but they are not
substantial.
93. Catalogue of the Demotic Papyri in the John Rylands Library, Manchester, III (Manchester, 1909), 185.
94. See now J. Cerny, "The Origin of the Name of the Month Tybi," Ann. Serv., XLIII (1943), 173-81.
95. Ostr. Deir el-Medineh No. 35:14 (published Cerny, Catalogue des ostraca hieratiques non litteraires de
Deir el-Medineh).
96. Ostr. Cairo Cat. 25598, 1 (published Cerny, Ostraca hidratiques, in Cat. gen.).
97. The lunar names are based on Pls. I - III and Fig. 16. Cerny's article, referred to above, should be con-
sulted for the names of the civil months. The most important sources are the calendar on the verso of the new
Cairo duplicate of the Calendar of Lucky and Unlucky Days of Sallier IV, and Brit. Mus. ostr. 5639a, both
Ramesside.
98. Edfou, VII, 7; IV, 8; Brugsch, Thesaurus, pp. 255: 40 and 266: 12.
99. Sethe, Zeitrechnung, p. 39. On the translation "emmer" cf. W. W. Struve, Mathematischer Papyrus des
staatlichen Museums der schinen Kiinste in Moskau (Berlin, 1930), pp. 62-63.
100. Beni Hasan, Vol. I, P1. XXIV, and the Ramesseum ceiling, Pairs of months, distinguished as "big" and
"little" etc., are a widespread phenomenon; their existence is due to a longer season being involved than can
easily be included in one lunation. Cf. Nilsson, Primitive Time Reckoning, pp. 224-25.
101. Cf. the table in Wilkinson, The Manners and Customs of the Ancient Egyptians, ed. Birch, II, 398-99.
102. Baedeker, Egypt (1929), p. xxiii. Records maintained by the Epigraphic Survey do not bear out this last
figure. In 1946-47, for example, the mean minima were December 450, January 410, February 440, and March 53
.
103. This was kindly brought to my attention by Professor W. F. Albright.
104. Langdon, Babylonian Menologies, pp. 133-35.
105. Georges Dossin, "Les archives 6conomiques du palais de Mari," Syria, XX (1939), 105.
106. Zeitrechnung, pp. 37-42.
107. Mittel, p. 6, nn. 8 and 9.
108. Op. cit., pp. 51-55. This statement will require modification, at least in part, if an interesting theory of
A. Varille's on the meaning of a passage in the Petrie stela concerning the obelisks which once stood before the
temple of Amon Re Montu, just north of the great temple of Karnak, proves to be correct. According to Varille
(Karnak I [Le Caire, 1943], p. 15 and n. 1), the passage in question, which tells of the erection of the obelisks
S~- oign\ "one on each road between which my father (Re) rises," has to do with the solstices.
He promises to discuss the point in detail later.
109. Botti-Peet, I1 Giornale della Necropoli di Tebe (Torino, 1928), Pl. 5, 2. Cf. also Gardiner, ZAS, XLIII
(1906), 138, 142.
110. At Dendera the references are numerous in the texts on the eastern stairway, whose reliefs depict the
77
oi.uchicago.edu
NOTES
procession to the roof on the occasion of Hathor's New Year festival. Cf. particularly Mariette, Denderah IV,
P1. 3, above Figs. 4 and 5; P1. 20, Fig. 44; and P1. 24 b, bottom. On I _t 1 in the calendar of Edfu (Edfou, V,
348, 5) was celebrated "the feast of Harsamtawi, lord of gJ-dy, in his beautiful feast of the birth of the sun disk."
111. Brugsch, Thesaurus, p. 365, 1; Duemichen, Altaegyptische Kalenderinschriften (Leipzig, 1866), P1. 105,
left, and P1. 112, 32-33.
112. Op. cit., p. 142.
113. As is well known, in A.D. 139 Sothis rose on I t 1 (Censorinus De die natali 21, 9-12).
114. Brugsch, Thesaurus, pp. 380 ff.; LD IV, 78a.
115. Reading hb wp rnpt dd tpyw
- c
, with dd as passive participle.
116. The latest cycle mentioned in the Carlsberg Papyrus began in A.D. 144; the following one would, of course,
begin in 169.
117. In A.D, 565 Olympiodorus wrote that the observation at Memphis determined the Sothic date for the whole
land. Cf. Sethe, Zeitrechnung, p. 309.
118. In the Edfu calendar (Edfou, V, 351, 10) after the rites of the death and burial of Osiris comes this entry
on I prt 1: "Feast of wp rnpt. Feast of Horus, [son of Osiris] and Isis. Accession of Horus of Belhdet, the son
of Re, beloved of mankind. Performing all ceremonies like I __t 1." Thus in the Ptolemaic period wp rnpt could
signify the opening of the reign of a king.
119. Cf. J. G. Milne, A History of Egypt under Roman Rule (3d ed.; London, 1924), p. 58. The news of the
accession of Pertinax to the throne in A.D. 193 took 65 days to travel from Rome to Alexandria. On its arrival
the prefect ordered a 15-day festival. Fifty-two days after the murder of Pertinax an official document was
dated with his name in Egypt. Cf., also, F. Hohmann, Zur Chronologie der Papyrusurkunden (Berlin, 1911), pp.
50 ff., who lists P. Tebt. 481, still dated to Antoninus Pius on April 25, A.D. 161.
120. Edfou, V, 351, 10.
121. Specifically from one of the years 398, 373, 348, 323, or 298 B.C.
122 .Cf. 102.
123. Wb., I, 437, 10.
124. Edfou, VI, 121, 9; cf. H. W. Fairman, "The Myth of Horus at Edfu," JEA, XXI (1935), 32.
125. Edfou, VI, 123, 2, and Fairman, op. cit., p. 33.
126. Edfou, V, 351, 5.
127. As in Edfou, V, 351, 6.
NOTES TO CHAPTER IV.
1. Chronologie, pp. 3-44.
2. The Sothic year was not exactly 365 1/4 days but was so close to that figure that to all intents and pur-
poses the Sothic year and the Julian year were one and the same in length throughout Egyptian dynastic history.
Those interested in more exact figures should consult Borchardt, Mittel, pp. 10 ff., and the literature cited there-
in. There they would discover that Borchardt rejected the cyclical calculation of the Sothic period and correctly
insisted that astronomical calculation must be used. His calculated "first certain date" was July 16, 4226 B.C.
3. "Die Bedeutungslosigkeit der 'Sothisperiode' fur die alteste agyptische Chronologie," Acta Orientalia,
XVII (1938), 169-95. He reaffirms his position in "The Origin of the Egyptian Calendar," JNES, I (1942), 396-
403.
4. A. Scharff, "Die Bedeutungslosigkeit des sogennanten altesten Datums der Weltgeschichte," Historische
Zeitschrift, CLXI (1940), 3-32. He accepts Neugebauer's theory but decides on the basis of present evidence
that the 365-day year need not have been inaugurated prior to ca. 2800 B.C. Scharff had earlier come to the con-
clusion that the calendar could not have been instituted in 4241 B.C. but began rather a whole Sothic period later,
in 2781 (2776 was his figure); cf. Grundziige der Agyptischen Vorgeschichte (Leipzig, 1927), pp. 54 ff.
5. JNES, I (1942), 401, n. 17.
6. Loc. cit.
7. "The Origin of the Ancient Egyptian Calendar," Proceedings of the American Philosophical Society,
78
oi.uchicago.edu
NOTES
LXXXIII (1940), 450. Neugebauer has replied to this criticism (JNES, I, 397, n. 3) but not, I feel, convincingly.
Averaging could hardly be as simple a process as he makes it out to be. A continuing operation from year to
year of addition and division would be necessary.
8. Op. cit., p. 462.
9. Urk., I, 25.
10. The figures given are the actual lengths of lunar years according to the Babylonian calendar of 310 to
286 B.C.; cf. Parker and Dubberstein, Babylonian Chronology from 626 B.C. to A.D. 45 (Chicago, 1942; 2d ed.,
1946), pp. 35-36. The empirical intercalation of the Egyptian calendar will have agreed with or very closely
approximated the Babylonian cycle, and the figures may be used with confidence.
11. Naturally the Egyptian did not use the decimal system for his averaging. One might imagine some such
simple process as this: Since 354 is the smallest number, use that as a base and total the days in excess of
that figure for the first 11 years. The total is 122. Divide 122 by 11, and the result is 11 with 1/11 remainder.
354 plus 11 equals 365.
12. On the origin of the 30-day month cf. Neugebauer, JNES, I, p. 400.
13. A later parallel, as we have seen (110-122), is the 25-year cycle, the first day of which begins both a
civil and a lunar month.
14. Since the rising of Sothis moves in the civil calendar one day every four years, 2937 results from 164
(41 times 4) added to 2773, and 2821 from 48 (12 times 4) added to 2773.
15. In the Middle Kingdom, at any rate, the year was so reckoned (Sethe, Zeitrechnung, p. 303). The five
epagomenal days were above and beyond the year.
16. A tradition recorded by P. Nigidius Figulus in the first century B.C.; cf. Sethe, Zeitrechnung, p. 310,
and Winlock, op. cit., p. 463.
17. Georges Daressy, Statues de divinitis (2 vols.; Le Caire, 1905-06), pp. 231-34, P1. XLVI.
18. Plate IV (Edfu) shows thirty-six of the forty-eight deities, omitting in each group the intrusive one.
Brugsch's numbering is wrong. The erased figure at the beginning should be numbered 1, and so on.
19. Generally speaking, the decanal lists of the Greco-Roman period reflect an earlier tradition than those
which immediately precede that period. Cf., e.g., Brugsch, Thesaurus, pp. 137-43, where the earlier lists may
be found, ending with that on the sarcophagus of Nectanebo. These lists in turn vary considerably from those of
the Middle Kingdom diagonal calendars, of which Brugsch was ignorant. The writer and Professor Otto Neuge-
bauer have in view a publication of all Egyptian astronomical texts, wherein the problem of the decans will be
studied in extenso.
20. Brugsch, Thesaurus, pp. 151-52.
21. Journal of entry, No. 41751. The preserved portion is 87 mm. wide and 70 mm. high. Cf. Georges
Daressy, "La Semaine des Egyptiens," Ann. Serv., X (1910), 181-82.
22. "La Semaine des Egyptiens," op. cit., pp. 21-23 and 180-82. The stela from Tell Basta which he dis-
cussed in his first article has seven registers of deities, the last of which presumably included Nos. 50-59 (see
Ahmed Bey Kamal, "Notes prises au cours des inspections," Ann. Serv., IX [1909], 191-92 and plate); and a
hollow cylinder of glazed clay, now in the Cairo Museum, which he discussed in his second article, has but Nos.
50-59 around it. The serpent, No. 51, omitted between Osiris and Horus was added after No. 59 by the ancient
craftsman.
23. The "weeks" would consist of 7, 8, 7, and 7 or 8 days, respectively. There can be little doubt that the
week of 7 days was a schematic fourth of the lunar month.
NOTES TO EXCURSUS A
1. Alan H. Gardiner, "Mesore as First Month of the Egyptian Year," ZAS, XLIII (1906), 136-44.
2. Suggested tentatively in ibid., p. 139, and strongly in The Tomb of Amenemhet (London, 1915), p. 97, n. 4.
The identification has the approval of Sethe, Zeitrechnung, p. 31, n. 2, and Weill, Bases ... de la chronologie
egyptienne, p. 117.
3. E. Meyer, Nachtrage zur aegyptischen Chronologie (Berlin, 1908), pp. 3-18.
79
oi.uchicago.edu
NOTES
4. In the calendar of Esna (Brugsch, Thesaurus, p. 382, 12).
5. Zeitrechnung, pp. 30-37. Weill (oP. cit., pp. 112-26) accepted all Sethe's conclusions and elaborated on
them.
6. Cf. Botti and Peet, Il Giornale della Necropoli di Tebe (Torino, 1928), p. 53, n. 7, and P1. 59, 1. 19.
7. Edfou, V, 350.
8. Ibid.
9. J. J. Tylor and F. L1. Griffith, The Tomb of Paheri at el Kab (London, 1894), P1. IV, and Urk., IV, 109.
10. Brugsch, Thesaurus, p. 380:5.
11. At Medinet Habu (MH, Vol. III, P1. 154: 686-704) are listed supplies for but one day. The only evidence
for a duration of the feast for more than one day is in a papyrus of the second century after Christ (P. Brussels
E 7535). The feasts of Hermes, beginning on Thoth 19, lasted seven days, but it is quite uncertain whether this
duration reflects ancient custom or simply a transplanted Greek one. Cf. Marcel Hombert and Claire Preaux,
"Les papyrus de la Fondation Egyptologique Reine Elisabeth," Chronique d'Egypte, XV (1940), 136 and 143.
12. This seems to be the explanation which Borchardt offered of the "Gardinerschen Phanomens"; cf. Mittel,
p. 24. But mere double-dating in the civil and later lunar calendars would never account for the feast of Renutet
falling, in every known instance from the 18th dynasty to the Roman period, on I smw 1. Moreover, the calendar
of Medinet Habu differentiates between feasts determined by the moon and feasts fixed in the civil year, and
there is nothing there to indicate that the feast of Renutet on I smw 1 was lunar.
13. Edfou, V, 348-60.
14. I have omitted the last four months of the calendar because their feasts have no bearing on the demonstra-
tion.
15. Ibid., p. 352, 14.
16. Ibid., p. 356, 22.
17. Brugsch, Thesaurus, p. 367: 20-21.
18. Ibid., p. 382: 11.
19. Ibid., p. 46; also my P1. V.
20. Mariette, Denderah, Vol. IV, P1. 77a, with some corrections from Brugsch, Thesaurus, pp. 325-28. A
collation of this important text is badly needed.
21. Edfou, V, 350, 9.
22. Cf. V. Loret, "Les fetes d'Osiris au mois de Khoiak," Recueil de travaux, III (1882), 43-57; IV (1883),
21-33; V (1884), 85-103; and H. Brugsch, "Das Osiris-Mysterium von Tentyra," ZAS, XIX (1881), 77-111.
23. It is my belief that the whole character of Egyptian kingship in its transmission from father to son will
be found to have, ultimately, a lunar explanation. The dying Horus is the waning moon. Horus dead becomes
Osiris, and the moon is invisible. The new crescent is the symbol both of the reborn Osiris as king of the dead
and of his son and successor Horus as king of the living. The ceremony of the Sed is also understandable on a
lunar basis. The king normally celebrated it for the first time in his 30th year. The 29 years he has reigned
may be compared to the 29 days of the lunar month from the day of new crescent, the second day, to day 30.
Like the moon, the king, Horus of the Living, has run his course. In his 30th year, pdntyw, he dies and is re-
born as the youthful Horus, the crescent moon, strong and vigorous.
24. This is in itself an indication of a lunar origin. On the importance of the first six days of the lunar
month as being the six parts of the eye of Horus, cf. H. Junker, "Die sechs Teile des Horusauges und der
'Sechste Tag,' " ZAS, XLVIII (1910), 101-6.
NOTES TO EXCURSUS B
1. E. Meyer (Chronologie, p. 36) considered tpy rnpt to mean "first day of the wandering year."
2. Thesaurus, pp. 213 and 1124.
3. Beitrdige zur Altesten Geschichte Agyptens, p. 136; Kommentar .. . Pyramidentexten, IV, 16.
4. "Horus the Beidetite," JEA, XXX (1944), 30.
5. "Jahre und Tage der Kr6nungs-Jubilien," ZAS, LXXII (1936), 52-59.
80
oi.uchicago.edu
NOTES
6. According to Borchardt, on the anniversary of his "coronation"; but see A. H. Gardiner, "Regnal Years
and Civil Calendar in Pharaonic Egypt," JEA, XXXI (1945), 24, who rightly insists on a distinction between
accession and coronation, the former of which took place immediately on the death of the king's predecessor.
In the 18th and following dynasties of the Empire, regnal years were reckoned from the day of accession.
7. Mond and Myers, Temples of Armant (London, 1940), P1. XCIII 1 and pp. 163 ff.
8. Ibid., P1. XCIII 3.
9. Op. cit., p. 30, n. 4.
10. Borchardt, op. cit., p. 53.
11. Ibid, pp. 53-54.
12. A. H. Gardiner, "The Goddess Nekhbet at the Jubilee Festival of Ramses III," ZAS, XLVIII (1910), 49.
Borchardt, op. cit., p. 54, quotes Spiegelberg's old and incorrect reading, I smw 28.
13. Gardiner, in ZAS, XLVIII, 48-49.
14. Borchardt, op. cit., p. 58. See also Robichon and Varille, Le Temple du scribe royal Amenhotep fils de
Hapou (Le Caire, 1936), P1. XXXV.
15. This important tomb has only recently been refound; cf. Ahmed Fakhry, "A Note on the Tomb of Kheruef
at Thebes," Ann. Serv., XLII (1943), 447-532.
16. Ibid., p. 478.
17. Beitrage, p. 136. Cf. Michel Malinine, "Calendrier "gyptien des jours fastes et n"fastes," Melanges
Maspero, I, 892-93, for a discussion of the meaning of hd t:. In the astronomical day it meant the period of light
before sunrise and is thus strictly a part of the preceding day. In the natural day of the people, who were up be-
fore sunrise, it meant simply "daybreak" of the new day.
18. J. A. Wilson, "Illuminating the Thrones at the Egyptian Jubilee," JAOS, LVI (1939), 294; LD III, 84a.
19. Mariette, Abydos, I, 51: 44-47 - Brugsch, Thesaurus, pp. 213, 1124.
20. So considered by Moret, Du caractere religieux de la royaut
6
pharaonique (Paris, 1902), p. 256, who trans-
lates: "... tu t'es lev
6
sur ton pavois de la fete sed, tel que R& au debut de l'ann6e."
21. G. Jequier, Le Monument Funeraire de Pepi II, Vol. II (Le Caire, 1938), P1. 50.
22. Belegstellen to Wb., II, 292, 4-5. The only bit of contradictory evidence is from one of the calendars of
Edfu, where IIII Lt 29 is called Hathor's "beautiful feast of nhb kiw," and the following day is denoted 'the second
day of the feast of this goddess' (Edfou, V, 350, 9-10).
23. Thesaurus, p. 395, 1126.
24. Wb., II, 291, 15 and Belegstellen; A. W. Shorter, "The God Nehebkau," JEA, XXI (1935), 47.
25. WB., II, 292, 4 and Belegstellen.
26. Edfou, V, 350, 10, and 399, 7.
27. This may be the conception behind the two quotations of Brugsch given above with reference to Harsamtawi.
28. Borchardt in ZAS, LXXII, p. 57, n. 5, believed that n1b kiw as a New Year's feast was probably the name
of the first day of the regnal year of Horus as king of Egypt.
NOTES TO EXCURSUS C
1. "Chronology of the Twelfth Dynasty," JNES, I (1942), 307-14. Giulio Farina, Il Papiro dei re restaurato
(Roma, 1938), p. 63, had challenged the date as being incorrectly ascribed to Sesostris III.
2. Op. cit., p. 310.
3. By a remarkable coincidence, the results reached by Wood, "The Kahun Papyrus and the Date of the
Twelfth Dynasty," BASOR, No. 99 (October, 1945), pp. 5-9, for the first year of Sesostris III and the beginning of
the dynasty are the same as mine. His two errors, the one in taking crescent visibility to be the starting point
of the month and the other in ascribing Berlin Pap. 10056 (see below) to the reign of Sesostris HI, cancel one
another exactly.
R. Weill's proposal to make the 12th dynasty contemporary with the Hyksos and reduce the period between
its end and the beginning of the 18th dynasty to a maximum of thirty years (cf. "Remise en position chronologique
et conditions historiques de la XIIe dynastie 4gyptienne," Journal asiatique, 1947, pp. 131-49, and "Le Synchro-
81
oi.uchicago.edu
NOTES
nisme 6gypto-babylonien du d6but du IIe millenaire et l'evolution presente de la chronologie babylonienne,"
Chronique d'Egypte, No. 41 [January, 1946], pp. 34-43) has been criticized by Jean Capart, "Remarques sur
l'article pr4ced4nt," Chronique d'Egypte, No. 41, pp. 44-45, and C. F. A. Schaeffer, "A propos de la chronologie
de la XIIe dynastie egyptienne et des Hyksos," Chronique d'Egyypte, No. 44 (July, 1947), pp. 225-29. To their ob-
jections may be added the evidence in this excursus of the complete agreement of the lunar data from the 12th
dynasty with the Sothic date of Sesostris III, which excludes a shift for the dynasty of some two hundred years.
4. Mittel, pp. 44-45.
5. There are others with varying degrees of uncertainty attached to the' readings; cf. ibid., pp. 36-46.
6. ZAS, XXXVII, 92-93; cf. also Mittel, pp. 7 and 29 ff.
7. Some time ago Dr. W. Erichsen collated the original for 0. Neugebauer, and the latter has kindly allowed
me to quote his remarks.
8. Gardiner and Erichsen agree on 19, while Borchardt read '
2 0
'. This it cannot be, since, as Gardiner
pointed out in a personal letter, the sign immediately before the 9, which itself is mainly lacuna, is clearly the
10 before units and not the first part of 20, which is always different in form; cf. also Gardiner, "Regnal Years
and Civil Calendar in Pharaonic Egypt," JEA, XXXI (1945), 22, n. 2.
9. Borchardt regarded 20 as uncertain, but it is perfectly clear.
10. The 10 is clear, but the units are in lacuna. Any other reading is excluded by the preceding and following
dates.
11. Chronologie, p. 52.
12. Pointed out with emphasis by E. Mahler, Etude sur le calendrier egyptien (Paris, 1907), p. 131.
13. "The Chronology of the Twelfth Dynasty," JEA, IX (1923), 199, quoted and approved by Edgerton, op. cit.
p. 310.
14. It cannot be entirely excluded as a possibility that here the exact meaning of this preposition is "down to,
but not including," as in Gardiner, Grammar, p. 135.
15. Berl. Mus., Pap. 10003 A, ii, 16-19, published in Moller, Hieratische Lesesticke, I, 18, and Gardiner,
Grammar, pp. 255-56.
16. Mittel, p. 45.
17. Gardiner, op. cit., pp. 21-23.
18. Borchardt had demonstrated (op. cit., pp. 31 and 45) that C cannot fit with D but does fit with A. As he
had already assigned D to Sesostris III, A had to be, therefore, Amenemhet III. Fortunately for his argument
as it stands, he failed to try to fit B with either.
19. Hypothesis I permits an agreement between a restored and a calculated 1111 t 18, while Hypothesis II
permits none.
20. Edgerton has raised the point (op. cit., p. 313) that, since we have no pre-Manethonian document spe-
cifically dated to Sesostris III in any year after his nineteenth, it is not absolutely impossible that the Turin
scribe may have reversed the order of Sesostris II and Sesostris III and his real opinion may have been that
Sesostris II reigned 30 + x years and Sesostris III but 19.
Now if Sesostris III's year 7 was 1870 + 6 years and his reign was 19 years long, years 30-31 of
Amenemhet III must have been 1828-1827 + 6 years. The only years which could possibly fit D would be 1827-
1826, and calculation shows that the fit is a bad one. Let us assume, on the other hand, that D belongs to Sesos-
tris II. As his reign might be from 31 to 39 years long, years 30-31 would fall between 1891 and 1871. The
only possible years for D are 1877-1876 and 1888-1887. Upon calculation neither one will be found to yield a
good fit. We may confidently conclude that the Turin scribe did not mistakenly reverse the reigns of Sesostris
II and III.
21. Borchardt had read Mktn s? ... -sub (ZAS, XXXVII, 93). Posener prefers Mkt 10I , to Mktn, pointing
out that Ranke lists no hieroglyphic examples of the latter and the cursive hieratic of the Middle Kingdom does
not distinguish between __ and
22. Griffith, Hieratic Papyri from Kahun and Gurob (London, 1897), Pls. X, XI.
23. The above discovery was communicated to me by Posener in a personal letter, and I am deeply indebted
82
oi.uchicago.edu
NOTES
to him for it. He commented further that a temple-scribe named Hr-m-s.f, the recipient of many of the Illahun
letters published by Scharff (ZAS, LIX, 20-51), had become an overseer sometime between years 18 and 24 of
Amenemhet III (op. cit., p. 22) and that if the IHr-m-s:.f named as temple-scribe in 1. 2 of Pap. 10056 were the
same man it would be necessary to consider whether the reading of "year 31" in 11. 2, 5, and 6 should not rather
be "year 21." But he admits that the hieratic does not lend itself to that reading, and it is sufficient to check
that hypothesis by means of the 25-year cycle to determine that, if the years involved were 20 and 21, the lunar
months could not be fitted with either A, B, or C, and we should have three groups of lunar dates 21 years or
higher, an impossibility for the kings involved. The obvious conclusion is that there was a second and later
temple-scribe with the common name of Hr-m-s .f.
24. Correct for Borchardt's 31-day month which ended It 20, but not for ours ending II lht 20.
25. M. G. Posener has made several valuable suggestions for which I am grateful. There can be little doubt
that t-sp 9 is the correct reading in line 1. The original must have been , with as in Kahun Pap.,
Griffith's Pls. XV, 13 and XXXIII 32. What follows in line 1 is quite uncertain but not important. Borchardt
(Mittel, p. 34, n. 5) gave the date as "Jahr 9, 29. (?) 10. (?) W.," but the photograph shows clearly that no other
month or day is possible.
26. Mittel, p. 29.
27. Edgerton, op. cit., p. 312.
28. Ibid.
29. Op. cit., p. 85.
83
oi.uchicago.edu
oi.uchicago.edu
PLATES
oi.uchicago.edu
oi.uchicago.edu
P~LATE I
-J
VC
liii- As'nioNOMI(AL ('FALL r. IS TfIF TOMBu OF SENMUT
T
f
oi.uchicago.edu
tII'lI 11
THE ASTRONOMICAL CEILI\G OF THE RAMIESSEUM (EASTERN HALF)
oi.uchicago.edu
P~LATE III
TE ASPONO(MI( AL CEfI LIN OF THEI!I H AMEAEUN (WETERN HIALF
oi.uchicago.edu
P~LATE IN'
a. ;* a. aa.
THE ASTRONOMrICAL FIEZE OF~ THE TEmPLE OF EDFU (FIRST HALF)
oi.uchicago.edu
PLATE V
THE ASTRONOMICAL FRIE/E OF THE TEMPLE OF EDFU (,S:E(OND HIALF)
oi.uchicago.edu
PLATE VI
At, 'PIIE CAIRO MCM11U-FRAGMIENT (A'rTUAL SIZE)
wp-
"N
B, CAIRO PAPYRUS 58065, RECTO
oi.uchicago.edu

Calendars and Astronomic Ceilings

Uploaded by

Copyright:

Available Formats

Calendars and Astronomic Ceilings

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Calendars and Astronomic Ceilings

Uploaded by

Copyright:

Available Formats

THE ORIENTAL INSTITUTE OF THE UNIVERSITY OF CHICAGO

STUDIES IN ANCIENT.ORIENTAL CIVILIZATION" NO. 26

. At Luxor the mean maximum is 760 in December, 740

Memphis (lat. 29.90) " " 9.5 to 8.8

Illahun (lat. 29.2

You might also like