CRISTIAN TOLSA

T IME OF B IRTH AND A SCENDANT IN THE P APYRUS H OROSCOPES

aus: Zeitschrift für Papyrologie und Epigraphik 204 (2017) 209–220

© Dr. Rudolf Habelt GmbH, Bonn

 209

TI M E OF B I RT H AND A SCEN DA N T IN THE PA PY RUS H OROSCOPES

It is well known that the ascendant played a capital role in Greek astrology. The degree of the zodiac rising
on the eastern horizon at the time of birth was indispensable for the determination of the lots, of the other
“centers” (the mid-heaven, the descendant, and the lower mid-heaven) and of the δωδεκάτροπος or twelve
houses, all of which providing a good part of the interpretational material at disposition of the ancient
astrologer. Also, it was the basis for most of the procedures used for the calculation of the length of life.
At the same time, the speed of the daily revolution of the heavens imposed great difficulties in the
determination of the rising point. This contradiction is the main matter of Sextus Empiricus’ essay against
the astrologers in Adversus mathematicos, where it is summarized in three objections: the problem of
ascertaining the exact moment a person is born (M V 55–67), the difficulty of observing the ascending
sign (68–85), and the unreliability of time reports (86–87) – the rest of the essay draws from more general
arguments against astrology. We can concede to the first of these objections, since one sign ascends on
average in only two hours, a circumstance that demands a precise definition of the time of birth; if the birth
takes place close to the boundary between two signs, it makes a difference whether we consider that the
child is born when the head is out, or when the whole body has been delivered. The second objection can
be dismissed, since Graeco-Roman astrologers did not use reports of direct observations of the skies, but
exclusively ephemerides and almanacs with planetary positions (including the sun and moon) computed
with various kinds of algorithms: Babylonian-style arithmetical schemes, and, later, Ptolemaic geometrical
models. The ascendant degree was in turn generally calculated from the position of the sun and the time of
birth, using a table of ascensions (rising times) for the latitude of the birth in question, modelling the une-
qual time employed by each of the signs to emerge from the horizon. Two kinds of tables were used here, as
well: Babylonian tables based on constant increases for the ascending times of the signs, and the Ptolemaic
table based on spherical trigonometry.
It thus seems that Sextus was misinformed about how astrologers produced the astronomical infor-
mation for their charts. Whereas he never mentions the use of astronomical tables, he alludes to the use
of water-clocks as assisting devices to keep track of the rising times in their observations (75–77), which
may have been his own interpretation of the mathematically-obtained tables of rising times. However, the
third of his objections, namely the problem associated with keeping track of time in antiquity, is valid.
Devices for time-keeping – sundials and klepsydras – were not always at hand in the ancient Greek and
Roman world, and even then, they were notoriously inaccurate.1 The latter is the reason why Ptolemy (2nd
c. A.D.) recommends the anaphorical clock in the Tetrabiblos (Tetr. III 1). Nevertheless, acknowledging
that such a sophisticated tool was rarely used, he offered an alternate astrological method for finding the
ascending degree given the ascending sign, using the longitude in the zodiac of the previous syzygy (new
or full moon) and choosing the degree occupied by the planet with the greatest number of affinities with
that longitude. There were a multitude of such methods described in the astrological literature, deducing
the degree through astrological procedures of varied complexity and assuming knowledge of the ascendant
sign, found astronomically from the approximate time.2
Other methods sought to correct the degree obtained from the table of ascensions, also with astrological
techniques. For example, Paulus of Alexandria (4th c. A.D.) records a method in which one had to pick the
degree closest to the given ascendant degree which had the same single-degree ruler (μονομοιρία) as the
sun or the moon, according to the daily or nocturnal nature of the birth (Isag. 22, p. 88 Boer). Valens (2nd
c. A.D.) reports a method by Thrasyllus (1st c. B.C./A.D.) in which the astrologer compares the two values
obtained from (1) adding the time-degrees from the sun to the moon computed with the table of ascensions
1 Apart from the fact that the minimum division was almost always the hour, it has been argued that an error of about 15
minutes could be expected in average. Cf. R. Hannah, Time in Antiquity, London/New York 2009, 106.
2 For a survey of these methods, see A. Bouché-Leclerq, L’astrologie grecque, Paris 1899, 386–390.
210 C. Tolsa

and (2) the multiplication of length of daylight by the hours of the nativity modulo 360, and adds or subtracts
a maximum of an hour to the reported time, depending on the difference between the two values (IX, 338).3
A method maybe by Valens himself has elements resembling the two kinds, since it first astrologically
determines a first ascendant degree given the sign and the degree of the sun, using an almost unique table
of “zodiacal places”,4 and then corrects this degree following another operation on the same table and the
degree of the moon.
Finally, other methods intended to astrologically determine the sign of the horoscope, presumably
when the hour was not available. Hephaestio in his 4th-century compilation (II, 90–91) and Valens (I, 18–19)
give several procedures for this.5
D. Baccani, in her continuation of O. Neugebauer and H. B. van Hoesen’s compilation of Greek horo-
scopes,6 asserts that from the available data, we are in no position to ascertain whether any of these meth-
ods found in the astrological literature were used on the papyrus horoscopes.7 I will attempt to assess this
view by surveying the horoscopes in the Baccani and Neugebauer/van Hoesen collections, along with the
Oxyrhynchus horoscopes later published by A. Jones.8
We have ca. 120 horoscopes in total, of which only 31 give the degrees within the signs. In turn, from
these 31, only 7 preserve enough data to check the calculation of the ascendant degree. Another 2 of them,
GH 46 and POxy 4279, show no degree for the ascendant, even if everything else is expressed in degrees – a
reflection of the difficulties perceived in its computation.

Determining the ascendant sign when client does not know the hour (GH 366, GH 373)

I shall begin with two plain horoscopes, GH 366 (PSI 22a) and GH 373 (PSI 24a), just providing the signs
of the planetary positions and of the ascendant.9 They belong to a set of three leaves containing ten horo-
scopes (including two pairs of duplicates) written by the same hand, together with a long multiplication
table and private accounts.10 These two horoscope reports are exceptional in that they do not give the hour
of birth, but just ὀψέ (“late”), a term generally indicating an hour close to the evening; as in English, it can
either mean late in the day or past the evening. The very fact that the sun is in different positions relatively
to the horizon in the two horoscopes, implying that the seasonal hour is different, implies that we are not
dealing with a synonym for the 12th hour (the last one) of the day, which could be the case in other instances,
3 I indicate the passages from Valens and from Hephaestio with book and Pingree’s page numbers. Valens does not pro-
vide any example of application of this method, and some aspects remain obscure, e.g. how the fraction to be added or subtract-
ed is to be calculated. One possibility is that the difference between the two quantities (which can be as high as 360 degrees)
is always taken in relation to 360, but this is far from certain. There seem to have existed many variants of this method: cf.
Val. I, 20–21; IX, 344; IX, 345–346. It is noteworthy that in all these variants Valens uses the name γνώμων to define the two
quantities. My guess is that the word was taken from the pointer of the hour in sundials. Since sundials were frequently inaccu-
rate, one was used to compare different sundials to get a sense of the hour, drawing a rough average; the term in Valens seems
to have the same connotation. See Hannah, Time in Antiquity, 134–135 for two sundials set side by side in Rome. Also, some
sundials showed the hour in more than two dials, such as the tower of the winds in Athens, an example of extreme virtuosity,
with dials on the eight faces plus a water-clock inside; cf. ibid., 164 n. 106.
4 See the only other known witness of this table in the papyrus analyzed in C. Tolsa, “New” Zodiacal Divisions in POxy
LXI 4277, ZPE 203 (2017), 192–198.
5 For Hephaestio’s methods, see S. Feraboli, Sulla μοῖρα ὡροσκοποῦσα (Heph. Teb. II 2), Quaderni Urbinati di Cultura
Classica, New Series 8 (1981), 157–160.
6 D. Baccani, Oroscopi greci. Documentazione papirologica, Messina 1991 (henceforth Ba); O. Neugebauer and H. B.
Van Hoesen, Greek Horoscopes, Philadelphia 1987 (henceforth GH). In the second of these collections the papyrus numbers
also indicate the date.
7 Ba, 79.
8 A. Jones, Astronomical Papyri from Oxyrhynchus, vols. I and II, Philadelphia 1999 (vol. LXI of the Oxyrhynchus
papyri).
9 In the illustrations, I only provide the elements of the horoscopes relevant to my analysis. The ascendant is always on the
left (center). Also, I write “rec.” in parentheses if the degrees or the hour have been reliably reconstructed.
10 See the schematic distribution of the space on the leaves in GH, 60.
 Time of Birth and Ascendant in the Papyrus Horoscopes 211

but with a certain indeterminate hour close to the evening.11 In our

case, the expression is embedded in both reports in the expression
for the date (τυβὶ [ιαʹ] ὀψὲ εἰς ιβʹ // τυβὶ ηʹ ὥρᾳ ὀψὲ εἰς θʹ), which
suggests that the births are understood to have taken place during
the night. Astrologers normally wrote down the day preceding and
the day after in this case: 11th and 12th in GH 366, 8th and 9th in
GH 373. Furthermore, as Neugebauer and van Hoesen pointed out,
the Lot of Fortune in GH 373 is computed according to the formula
for a night birth,12 and the position of the ascendant marks an hour
past the setting point in GH 366.
The indetermination of the time in which the person was born
would of course have been an important problem for the computa-
tion of the horoscope. Therefore, it is very probable that astrologer
resorted in this situation to some of the astrological methods availa-
ble to compute the ascendant sign without the hour. Valens provides
two methods of this kind, and Hephaestio another three (of which
one is attributed to Anubio). It is interesting that Valens and Hephaes-
tio each pretends that all the methods give the same result (Heph.
II, 91.15 ἄλλην καλλίστην μέθοδον καὶ αὐτὴν συμφωνοῦσαν),
the text in Valens even involving an intended miscalculation in the
example of the second method which makes it apparently consistent
with the first.13 One method using the so-called δωδεκατημόριον
of the degree of the sun appears with a small variant in the accounts
of the two writers;14 it is probably this one, the least cumbersome
of all, which the author of GH 366 and GH 373 is applying. In Valens it is the first method, in Hephaestio
the third:
Val. I, 18: ⟨Μ⟩αθὼν ἀκριβῶς πόσων μοιρῶν ἐστιν ὁ Ἥλιος ἐπὶ γενέσεως, ἰδὲ ποῦ τὸ
δωδεκατημόριον ἐκπίπτει· καὶ οὗ ἂν ἐκπέσῃ τούτου τὸ εὐώνυμον τρίγωνον ὡροσκοπήσει
ἢ τὰ ὁμοιόπτωτα ζῴδια, οἷον ἀρρενικὰ ἢ θηλυκά, νοοῦντός σου τὴν διαφορὰν νυκτὸς
ἢ ἡμέρας. οἷον ἔστω Ἥλιος Ὑδροχόου μοίρᾳ κβʹ· τοῦτο τὸ δωδεκατημόριον κατέληξε
Σκορπίῳ· τούτου τὸ εὐώνυμον τρίγωνόν ἐστιν Ἰχθύες. εἰ οὖν ἡμερινὴ ἦν ἡ γένεσις, ἔδει τοὺς
Ἰχθύας ὡροσκοπεῖν ἢ Ταῦρον ἢ Καρκίνον· εἰ δὲ νυκτερινή, τὰ τούτων διάμετρα· Παρθένος
ὡροσκοπήσει κατὰ τὴν πρώτην ὥραν.
Discovering exactly at how many degrees the sun is at birth, find where the dodekatemorion
falls, and the left triangle of the place where it falls will be rising, or the equivalent signs –
either masculine or feminine – having in mind the difference between night and day. Take, for
example, the sun in Aquarius 22; its dodekatemorion falls in Scorpio. The left triangle of this
is Pisces. Then, if the birth is diurnal, the ascendant is necessarily Pisces, Taurus, or Cancer; if
it is nocturnal, the diametral signs. In the first hour [of the night] the ascendant is Virgo.

11 The expression is used in relation with the hour in OKrok 29, ὥραν ιβʹ ὀψέ, indicating the 12th hour of the day, in the
evening. Cf. S. Remijsen, The Postal Service and the Hour as a Unit of Time in Antiquity, Historia 56 (2007), 127–140, at
137. For the use for an hour of the night, maybe less common, cf. Pl. Crat. 433a (cited in LSJ, s.v. ὀψέ), οἱ ἐν Αἰγίνῃ νύκτωρ
περιιόντες ὀψὲ ὁδοῦ.
12 GH, 71.
13 The second method consists, for night births, in adding the rising times of the sign of the moon (Scorpio in the example)
to the degrees of the sun, and counting this off from the sign of the sun giving one degree to each sign. The resulting sign should
be Aquarius, but the text says it is Virgo, just like the ascendant found with the first method (Val. I, 18).
14 The δωδεκατημόριον was obtained from multiplying the degrees by 12 and adding the resulting quantity of degrees
to the actual position.
212 C. Tolsa

Heph. II, 91: προδεδόσθαι χρὴ τὸ ἐν νυκτὶ τετέχθαι τὸν ζητούμενον ἢ ἐν ἡμέρᾳ, καὶ τούτου
δεδομένου σκόπει τὴν τοῦ Ἡλίου μοῖραν καὶ ταύτῃ ἀεὶ πρόσθες μοίρας ⟨β⟩ ⟨𐅶⟩ σκεπτόμενος
ἀπὸ Θὼθ ἕως Μεχείρ, ἀπὸ δὲ Φαμενὼθ ἕως Μεσωρεὶ πάλιν ἀεὶ ἄφελε μοίρας ⟨β⟩ 𐅶. καὶ τὴν
γινομένην οὕτω ποσότητα τῶν μοιρῶν τοῦ Ἡλίου δωδεκάκις ποιήσας ἔκβαλε ἀπὸ τῆς τοῦ
Ἡλίου μοίρας, καὶ εἰς ὃ ἂν ἐκπέσῃ ζῴδιον ἐκεῖνο ὡροσκοπεῖ ἢ τὸ τούτου τρίγωνον.
First, it must be known whether the concerned person is born in the night or in the day, and this
being given, inspect the degree of the sun and add to this always 2 1/2 degrees between Thoth
and Mechir, but between Phamenoth and Mesore always take 2 1/2. Then multiply the resulting
quantity of solar degrees by 12 and count off from the degree of the sun, and the place where
this falls will be the ascendant, or its triangle.
It seems plausible that the version in Valens, making it necessary to look at the left triangle (i.e. four signs
ahead) after finding the δωδεκατημόριον, was an evolution – or a misunderstanding, since it is superfluous
to the procedure as given here15 – from Hephaestio’s suggestion to look at the triangle in case the result-
ing ascendant sign is not compatible with the daily or nocturnal nature of the horoscope. On the other
hand, Hephaestio’s rule of adding 2 1/2 degrees between Thoth and Mechir and subtracting this between
Phamenoth and Mesore was surely designed to adjust the degrees of the sun when these were computed as
days from the beginning of the year,16 and therefore this rule did not have to be applied if the degrees had
been determined accurately. Both authors note that the method is useful if we know at least whether the
birth is by day or by night, but we are in an even better position in our horoscopes. The method applied to
our two births would indeed directly give the two ascendant signs in the text, using the computed degree
of the sun:17
sun δωδεκατημόριον
GH 366 Capricorn 17 Capricorn 17 + 204 = Cancer 17 + 24 = Leo 11
GH 373 Capricorn 15 Capricorn 15 + 180 = Cancer 15

Since with this result the hour comes in both cases very close to the evening, it would not have been neces-
sary to modify the ascendant by placing it in the left triangle.

Determining the ascendant degree in horoscopes with degrees

Let us now turn to the 31 horoscopes that show the degrees within the signs, and particularly to the 7 where
the indication of the hour is extant, along with the position of the sun (or its reliable reconstruction) and the
ascendant degree.18 The latter number is so small because most of the horoscopes including the degrees
(called “deluxe” by Jones) append detailed information after each of the longitudes, spreading the data
over a long stretch of papyrus, and making it more improbable that all the longitudes were preserved after
fragmentation – we should remember that all Oxyrhynchus papyri were discards. But even from the study
of these few horoscopes we learn that several approaches were used for the determination of the ascendant
point.
15 It is superfluous because Valens tells us to consider all the signs of the same gender (masculine or feminine). Since leap-
ing to the left triangle does not change the gender, which alternates in the order of the signs, the candidates will be the same.
16 It is easy to see from the longitudes of the sun at the beginning of the months — see e.g. the table in Jones, Astronomi-
cal Papyri I, 349 – that, by the month of Mechir, 3 degrees beyond the calculation using 30 degrees per month have accumulat-
ed; by Mesore, the slower motion of the sun has made up for these 3 degrees, and we must further subtract another 3 degrees.
A similar rule of thumb, counting from Thoth 1 but without the adjustment, is found precisely among Valens’ methods for the
determination of the ascendant (I, 19).
17 I use GH’s value, even if it reflects modern tropical longitudes; longitudes need to be adapted to the sidereal longitudes
used by the ancient astrologers, but the formula to that effect in Jones, Astronomical Papiry I, 343 gives just a third of a degree
less for the longitudes of A.D. 366, and similarly for A.D. 373.
18 In POxy 4275 Jones transcribes the ascendant, but this must be an editorial blunder, since the horoscope is obviously
cut off before the annotation of the ascendant, as can be seen from the image in the online database.
 Time of Birth and Ascendant in the Papyrus Horoscopes 213

(a) End of the hour (GH 95, Ba 14)

GH 95 (PLond 98) is a deluxe horoscope explicit about the fact that
the ascendant is fixed “in full correspondence between the table of
rising times and the klepsydra” (59–61: δι’ ὅλην τὴν ⟨ἀν⟩αλογίαν
ἐκ τοῦ ἀναφορικοῦ πρὸς κλεψύδραν). The editors express their
puzzlement at this remark,19 but it probably just means that preci-
sion to the degree was used for the computation, in contrast with the
usual practice with plain horoscopes (see below). Using a table of
rising times computed with system A,20 and the degree of the sun
as reconstructed by the editors,21 we discover that the agreement
is perfect if we assume that six complete hours have elapsed since
sunrise. It is unfortunate that the beginning is lost, since we do not
know whether the text gave a precise indication such as “6th hour
completed”, or just “6th hour”.

In Ba 14 (POxy XLVI 3298), the degree of the sun is not extant,

but it can be reconstructed by using modern tables and the formula
transforming to sidereal longitudes (see note 17 above), with a result
of Taurus 27. The indication of the hour is, according to Baccani’s
reasonable interpretation of the smeared letters after the number,
“7th hour end” (ὥρᾳ ζʹ ἔσχατα).22 Computation with the table of
rising times according to system B gives Pisces 4, very close to the
value Pisces 6 given in the text. By 219 A.D., the year of the horo-
scope, we can assume that these systems were still used.23

(b) Beginning of the hour (GH 260, GH 137a/b)

In GH 260 (POxy XII 1476) the hour is oddly given as “10th hour
completed and 2 degrees” (ὥρα ιʹ πεπληρωμένη μοιρῶν βʹ). Work-
ing with Ptolemy’s table of rising times from the Handy Tables we
reach Aquarius 26 at the end of the 10th hour, and the two extra
degrees give the exact result. Neugebauer and van Hoesen interpret
them as time-degrees, but we should probably expect χρόνοι instead
for denoting degrees on the equator.24

19 GH, 35.
20 I computed tables for the rising times according to systems A and B for the clime of Alexandria, with a simple linear
interpolation method for the degrees, as Valens explains in VIII, 290–291. For a simple and clear explanation of the Greek
version of the Babylonian arithmetical systems A and B for the rising times, see GH, 3–5.
21 GH, 35.
22 Ba, 152.
23 The ascendant using Ptolemy’s Handy Tables would be Aquarius 28. Cf. the critical edition R. Mercier, Ptolemaiou
Procheiroi Kanones. Ptolemy’s Handy Tables Vol. 1B: Tables A1–A2. Transcription and Commentary, Leuven 2011.
24 The result is however the same with this precision. Cf. M. Gansten, Balbillus and the Μethod of aphesis, GRBS 52
(2012), 587–602, at 601 for the same confusion between degrees along the zodiac (μοῖραι) and equinoctial degrees along the
equator (χρόνοι) in Neugebauer and van Hoesen’s translations.
214 C. Tolsa

In GH 137a/b (two duplicates, PLouvre N 2342 and PLond 110) the

hour is indicated as “first hours of the day, at the beginning” (ὥρας
αʹ τῆς ἡμέρας ἀρχάς). The sun’s position is given as Sagittarius
13;23, and the ascendant as Sagittarius 15. Therefore, it seems again
that two degrees have been added after the stroke of the hour. May-
be this practice was customary when the client indicated a birth at
the beginning of the hour.

(c) Neither beginning nor end: the moon-phase method? (POxy 4274, GH 81, GH 338)
This is perhaps the most interesting part of the analysis. For these three horoscopes, an astrological meth-
od using the phase of the moon could have been applied to determine the exact time within the hour.
This method appears in various passages of Valens’ book VIII, both explained and exemplified in several
case-studies as a preliminary to a calculation of the length of life. Even if the evidence is inconclusive due
to the astrologers’ silence on their procedure, in each of these horoscopes a plausible case can be made on
the basis of re-computation and other hints.

POxy 4274 contains two horoscopes, for the years 480 and 503
respectively. In the second one the ascendant is not preserved, so
we will concentrate on the first one. Here, the day and month are
unfortunately not preserved, but as Jones shows, only two consecu-
tive dates are possible for the recorded planetary positions.25 Using
Ptolemy’s table of ascensions, and the degree of the sun as calculat-
ed by Jones,26 the closest we can reach to the given ascendant, Cap-
ricorn 6, at the end of a certain hour (not preserved) is Capricorn
12 at the end of the 6th hour of the day, and Capricorn 13 for the
next day at the same time. It is theoretically possible that the hour
was indicated as “half past five”, which would have approximately
yielded this degree. I however favor an alternate hypothesis. Apart
from the fact that no other horoscope gives the half-hour,27 I will adduce that the other horoscope in the
papyrus shows data suggesting the use of Valens’ moon-phase method. An additional clue indicating that
these two horoscopes could have used this method is the unique circumstance that both originally showed
the degrees of the sun and the moon – precisely necessary for the procedure – but not those of the planets.
After the date, the second horoscope provides the day “6th according to the moon” (κατὰ σελήνην
ς). This evidently marks the days elapsed since new moon, which are actually about five and a half. An
indication of this kind is always present in Valens’ examples using the aforementioned method (VIII, 304):
Οὐεσπασιανοῦ ἔτος ζʹ, Ἐπιφὶ κεʹ εἰς τὴν κϛʹ, ὥρα νυκτερινὴ γʹ· κλίμα γʹ. Ἥλιος Καρκίνου
κζ μγʹ, Σελήνη Ἰχθύων ιβ νβʹ, πανσέληνος Ἐπιφὶ κβʹ, ὥρα ἡμερινὴ γʹ, Αἰγοκέρωτος κ⟨δʹ⟩.
ἀπὸ πανσεληνιακῆς ἡμέρας τε καὶ ὥρας ἐπὶ τὴν γενεθλιακὴν ἡμέραν τε καὶ ὥραν γίνονται
ἡμέραι γ ὧραι ιβ, αἵπερ εἰσὶ τοῦ ἀπὸ πανσελήνου δρόμου ἐπὶ σύνοδον (τουτέστι τῶν ιε)
εʹ λʹ. τούτων ἀφεῖλον ἀπὸ τοῦ μεγέθους τοῦ παρακειμένου τῇ τοῦ Αἰγοκέρωτος μοίρᾳ κʹ,
25 Jones, Astronomical Papyri I, 280.
26 Jones already uses sidereal longitudes, so we can expect high accuracy. His computation of the degree of the sun for
the second of the horoscopes on the papyrus gives exactly the degree in the text.
27 The measure of the half-hour was a rarity in normal social contexts, which is probably what the time reports in the
horoscopes reflect. Only a couple of surviving sundials show lines for this fraction: cf. Hannah, Time in Antiquity, 125. It is
however attested once as having been used by couriers: cf. Remijsen, The Postal Service, 138.
 Time of Birth and Ascendant in the Papyrus Horoscopes 215

ὅπερ ἐστὶ ιβ κʹ· καὶ λοιπαὶ γίνονται θ ιβʹ. τοῦτο μέρος ὥρας ἔσται. ἐψήφισα οὖν ὥρας β
καὶ προσέθηκα τὸ μέρος ⟨καὶ⟩ ἐγκλίματος τζ· γίνονται τμ νεʹ. ταύτας εὗρον ἐν τῷ ἐγκλίματι
περὶ τὴν κθʹ τοῦ Ὑδροχόου.
Vespasian 7th year, Epiphi 25th to 26th, 3rd hour of the night, clime 3. Sun in Cancer 27;43, moon
in Pisces 12;52, full moon was on Epiphi 22nd, 3rd hour of day, in Capricorn 24. From the day
and hour of the full moon to the day and hour of birth are 3 days and 12 hours, which is 7/30
of the period from full moon to new moon, this is, of the 15. I subtract this from the quantity
entered in Capricorn 20,28 which is 12;20, and the result is 9;12.29 This is the fraction of the
hour. I computed then 2 hours and added the fraction, plus the accumulated rising time, 307;
the result is 340;55. I found these in the accumulated rising times around Aquarius 29.
Despite a couple of minor mistakes, Valens’ procedure is clear. He first finds the proportion of days elapsed
since syzygy (new or full moon) in relation to the days of half lunar month (15). Then, instead of taking the
ascendant point corresponding to the end of the third hour, he takes it at the end of the second hour, and
adds a fraction of the hour corresponding to the complementary proportion (in our case, 23/30). This is
equivalent to subtracting the fraction of hour (7/30) from the end of the hour. In brief, he uses the semi-luna-
tion as a clock, so that if the birth falls in new or full moon, the hour given is considered as completed, but
as we advance in days from that point, we recede proportionally along the hour. To put a simpler example,
if seven days and a half have elapsed, and the indication is the 11th hour, the resulting ascendant will cor-
respond to half past ten. Valens explains the method in similar terms long before using it, at the beginning
of the book (VIII, 297).
Let us apply the method to the first horoscope. As it turns out, we are at the same point in the phase of
the moon as in the second horoscope on the papyrus, about five and a half days after conjunction, on the
first possible day; and six and a half on the second possible day. Assuming the latter date, and taking the
corresponding proportion of the 6th hour, the ascendant goes back to Capricorn 6, which is exactly what the
text shows.

The detailed horoscope GH 81 (PLond 130), written in 9 narrow

columns, is another case in point. Here the main data are extant:
the hour (3rd of night), the degree of the sun (Aries 14;3), and the
degree of the ascendant (Scorpio 18). Using the table of ascensions
for system A, the ascendant degree would be Scorpio 19;30, whereas
system B gives a worse approximation (Scorpio 20). Ptolemy’s table
gives Scorpio 21, but because of the date in the first century we are
surely dealing with one of the arithmetical systems.
The number could of course be the result of a mistake in the
computation, but there seems to be evidence that the astrologer used
the moon’s phase for another calculation; there is a good chance,
then, that he used it for the ascendant degree as well. First, however,
it is necessary to check if the moon-phase method gives the ascendant in the text. Considering the approx-
imately two days that have elapsed since new moon – the position of the moon according to the text is
Taurus 13 – the ascendant would fall indeed on Scorpio 18.
Now, in relation to the use of the moon’s phase for another computation in this horoscope, we have
the value for the duration of the pregnancy – 276 days – in a fragment recovered from the IX column
(ὁ σπόριμος χρόνος ἡμέρων σος). The motivation behind this calculation is that the length of the pregnancy,
calculated astrologically, was sometimes used by ancient astrologers to check the native’s chart at the time of
conception. As Neugebauer and van Hoesen point out,30 the computation does not agree with the procedure
28 It should be Capricorn 27 (the opposite of the degree of the sun).
29 Here again is a mistake: Valens just uses the integer part of the 12;20.
30 GH, 28.
216 C. Tolsa

as found in Valens (I, 49–50). Valens tells us to count the signs from the moon to the ascendant if the moon
is above the horizon, and subtract this number multiplied by 2 1/2 from the middle value of 273 days; or to
count the signs from the ascendant to the moon if the moon is under the horizon, adding the number mul-
tiplied by 2 1/2 to the middle value in that case.31 Consequently, if the moon was in the ascendant sign, the
days of pregnancy were 273. The procedure must have been common, because it appears in Hephaestio (II,
85) with only slight variations: instead of counting the signs and multiplying by 2 1/2, we count the degrees
and divide them by 13, the mean daily course of the moon; this is very close to Valens’ method, since the 30
degrees of a sign divided by 13 give 2.3, of which the 2 1/2 that Valens allots to each sign must be an approx-
imation. The other slight difference is that Hephaestio’s middle value for the pregnancy is 273 1/3 days.32
In our case, the moon is in the descendant, and thus the days of pregnancy would be very far from the
middle term in both versions of the method, whereas the number of the text assumes that the moon is just
about one sign ahead of the ascendant. If, however, the astrologer had used the position of the preceding
conjunction instead of the ascendant as the middle value, the number of days would have been 276 in any
of the two ways of calculating the distance, Valens’ and Hephaestio’s (approximating to the closest integer
in the latter version). Since the astrologer does not disclose his method, we cannot discard other possibil-
ities, but this is the only plausible explanation I can find for the number in the text.33 Having in mind that
the procedure, as found in these two authors, unnaturally measures the elongation of the moon from the
ascendant in days – unnaturally because the ascendant is not a fixed point from which the moon moves – it
is not difficult to think that our astrologer could have instead counted the days since conjunction, out of
carelessness or because it was a parallel version of the method.

There are grounds to believe that GH 338 (PSI 23a), the first and
most detailed of the set containing the horoscopes GH 366 and 373
seen above, also used an astrological method based on the moon
for the determination of the ascendant. This would not be surpris-
ing, given that in those two horoscopes an astrological method was
applied to the determination of the ascendant sign.
It is unfortunate that Hermesion, the native in GH 338 and prob-
ably the scribe of the whole set, wrote down the minutes but forgot
to provide the degrees of the sun and the moon in his horoscope.
These longitudes were however narrowed down by Neugebauer and
van Hoesen, thanks to the indication of the terms and the Lot of
Fortune, to Capricorn 0–5 and Scorpio 14–19 respectively.34 The
two longitudes are tied in the sense that if the sun is in Capricorn 5, the moon is in Scorpio 19, and so on.
According to modern computation (which needs almost no adjustment here because of the date) the sun
should be around Capricorn 4. The text however shows one degree less than modern computation for Sat-
urn, Jupiter, and Mars, so I will assume that the longitude computed by the astrologer was Capricorn 3. At
the end of the 3rd hour of the night, the ascendant would have been in Leo 18, whereas the text gives Leo
13. Even with the sun at Capricorn 1, which is the furthest it can go back taking the text into account, the
ascendant would be Leo 16, still too high.35
Therefore, we are in a situation where some method could have been used for determining an ascend-
ant degree between the one reached at the end of the hour and the one at the beginning. As it turns out,
31 K. Frommhold, Die Bedeutung und Berechnung der Empfängnis in der Astrologie der Antike, Münster 2004, 88–89.
32 Frommhold, Die Bedeutung und Berechnung, 131–154.
33 Neugebauer and van Hoesen mention that the third day after birth was considered the critical day according to a certain
doctrine, so that these three days could have been added to the middle value of 273; cf. GH, 28. This leaves however unex-
plained why the astrologer chose the middle value for his client.
34 GH, 67.
35 For this computation I have used Ptolemy’s table, but the results with tables for systems A and B are even higher, and
at this date it is reasonable to assume that the Ptolemaic table was used.
 Time of Birth and Ascendant in the Papyrus Horoscopes 217

Valens’ method of the moon would not work here as he applies it. However, if instead of subtracting the
proportion of days since syzygy to half cycle from the end of the hour, we add it to the beginning of the
hour, we obtain an ascendant degree very close to the figure in the text. About eleven days have elapsed
since full moon; adding the corresponding 13 time-degrees to the beginning of the third hour, we reach the
ascendant at Leo 14, just one degree above that of the text.
This could be a system alternative to that of Valens, using the whole cycle of the moon. Let us compare
them schematically: in every case the quantity to be subtracted or added is the number of days since the
beginning of the half cycle, divided by 15 and multiplied by the hourly time-degrees.
Half-month Valens GH 338 (?)
new moon to full moon subtract from end of hour subtract from end of hour
full moon to new moon subtract from end of hour add to beginning of hour

With the second method, one theoretical ambiguity in Valens’ method would be solved, namely the case of
a full-moon (sc. new-moon) birth. Indeed, in Valens’ system if we use the first half cycle, from new moon to
full moon, we obtain the beginning (sc. end) of the hour, but using the second half cycle the ascendant will
correspond to the end (sc. beginning) of the hour. Valens solves this by directly defining the hour at new
or full moon as full (VIII, 284). On the other hand, the second method would give the same result for full-
moon (sc. new-moon) using both sides of the definition, the beginning (sc. end) of the hour, since it is both
15 days (sc. 0) from new moon or 0 days (sc. 15) from full moon. It is even possible that the method used
by Valens was a corruption of this hypothetical version. We actually do not know if the two other analyzed
horoscopes used Valens’ version or this one, because the preceding syzygy was in both cases a new moon,
in which case the procedure would be the same.
My claim that GH 338 could have used the moon-phase method (in this version) gains force from the
fact that the astrologer indicated the “preceding ἀπόκρουσις” (17–18). Neugebauer and van Hoesen must
be right in the restoration of the Greek word, but they fail to interpret it correctly, referring to a doctrine
in Valens (V, 205–206) where the moon is said to incline (προσνεύει) to one sign or another depending
on where it is in its phase; the editors understand that ἀποκρούειν means “in the sign opposite to conjunc-
tion”,36 but in fact the verb refers to something as simple as “waning” (cf. LSJ, s.v. ἀποκρούειν). The previ-
ous “waning” then surely means the previous full moon, that is, the last time when the moon began to wane.
Neugebauer and van Hoesen were probably tricked by Valens’ example, which just gives the signs where
the moon is, not the degrees. The value of the papyrus (Gemini 2;29) adds more confusion to it, because it
contains an error. The actual previous full moon must have fallen about 20 degrees ahead, in Gemini 23 –
so maybe the scribe forgot to add the sign for 20 (κ), which would not be surprising given his fail to give
the degrees of the sun and the moon. That this was the intended longitude can be confirmed by the data of
the decan and the terms corresponding to this point, provided in the last lines, but here still more confusion
appears. Hermesion, the scribe, was obviously not the astrologer himself, because, as Neugebauer and van
Hoesen point out, he seems to have copied in the same column what must have been two different columns
with matching lines in the original.37 Lines 19–30 give the sign ruler, the ruler of the terms, and the decan
for the positions given earlier in lines 5–18. He made a mistake, however, forgetting to copy this extra data
for the Lot of Daimon, surely because its sign ruler, lord of terms and decan were the same as those of the
Lot of Fortune, coming just before (the positions are Virgo 27 and Gemini 20). Therefore, line 30 refers
to the place of the full moon, and the third decan indicated there implies a longitude between 20 and 30
degrees. It is now clear that the previous ἀπόκρουσις indicated the moment of the previous full moon.
In conclusion, 3 out of 7 horoscopes possibly used the moon-phase method for the determination of
the precise moment within the hour, 2 used the end-point, and 2 added two degrees to the end of the previ-
ous hour to signify the beginning of the next. Thus, in our admittedly small sample, we can say that more

36 GH, 66.
37 GH, ibid.
218 C. Tolsa

frequently than not an astrological method for the determination of the ascendant was probably used. It is
impossible to know whether this method or others were applied to the rest of the 31 horoscopes which show
degrees, but it might be significant that the fragmentary POxy 4282 mentions the days of the preceding and
the following syzygy.
It is important to underline that the moon-phase method was not exactly a correcting method in the
same sense as all other known astrological methods for the determination of the ascendant degree. Unlike
them, it aimed at determining the exact moment within the given hour, whereas the other procedures,
including those that chose a degree given the ascendant sign, and those that corrected the ascendant degree
obtained from the table of ascensions, could always result in an ascendant point corresponding to a time
that contradicted the report given by the customer. On the other hand, the moon-phase method never con-
tradicted the report; it just sought more precision, when there was no indication of when in the hour the
birth had taken place. Maybe it is significant that in our small sample this method was applied whenever
there was no such indication. Also, we should not underestimate the fact that it was relatively easy to
apply. Remember that in the case of the astrological methods used to determine the ascendant sign for an
unknown hour, it was also the one that was easiest to apply which we found in the papyri. Simplicity and
respect for the information provided by the customer could therefore have been crucial factors limiting or
facilitating the diffusion and use of these methods. This must have weighed much more, in the mind of
the astrologers, than the possibility of receiving an erroneous report of the time of birth which “should” be
corrected astrologically.

Determining the ascendant sign in plain horoscopes

The last question that I want to address is the determination of the ascendant sign in the most usual kind
of horoscopes, those that only note the planetary positions and the ascendant by sign, without the degrees.
It is accepted that these horoscopes used sign-entry tables, providing just the day when the planets entered
a new sign.38 Even if the preserved sign-entry almanacs do not include the sun, it is likely that when pro-
ducing these horoscopes astrologers made use of similar tables for the solar longitude, too; only we lack the
papyrological witnesses, probably because tables for the longitude of the sun could be re-used every year
of the civil Alexandrian calendar – with a minor adjustment depending on the position within the 4-year
cycle – so that they would not have been frequently discarded.
I have analyzed 21 horoscopes from the three collections, chosen with the criteria that they provide the
hour, the sign of the ascendant and can be securely dated after 250, for the reason that more accuracy and
a better agreement with our re-computation – using Ptolemy’s table – can be expected for these later times.
I have classified them in the table below according to the relation between the given ascendant and the
ascendant computed for the end of the given hour. It is perhaps surprising that out of these 21 horoscopes,
only 6 indicate an ascendant coinciding with the sign rising at the end of the given hour. From the remain-
ing 15, 10 choose the previous sign, and 5 the following sign.
I envisage two possible rough methods leading to these results, whose reconstructed outcomes have
been given in the last two columns. The first represents a method based on the rising times for whole signs.
The longitude of the sun would have been taken at the beginning of the sign. Thus, only a very simple table
of rising times would have been needed, just indicating the time-degrees in which each sign ascended.
Sometimes, when the longitude was past about 15 degrees in the sign (which can always be roughly esti-
mated from the date and a sign-entry table), the beginning of the next sign would have been chosen instead.
This often led to a wrong ascendant sign, past the end of the hour. Indeed, the degree at which one should
take the beginning of the next sign to obtain the correct ascendant sign depends on both the sign of the sun
and the given hour, and can be both higher and lower than 15. I have computed the possible cases where the
next sign could have been taken and written the result alongside the value using the current sign.

38 Jones, Astronomical Papyri I, 176.

 Time of Birth and Ascendant in the Papyrus Horoscopes 219

In this method, the safest option would have been to take the beginning of the current sign of the sun.
Thus, if the computed sign turned out to precede the one at the end of the hour, it could almost always be
understood that the hour was not completed. It could still happen that the ascendant obtained by using the
beginning of the sign fell one sign before the sign corresponding to the beginning of the hour, but this does
not happen often (only once in our examples).39 Therefore, in most cases this simplification functioned like
a randomizer selecting a time within the interval of the given hour, similarly to the moon-phase method.
As is clear from the table, all the given ascendant signs in these horoscopes can be explained as the
result of using this method. However, I have also checked whether the simpler method, allotting 15 degrees
of the zodiac to each hour (attested in Manil. III 218–224/483–502, Paulus Alex. 29 p. 80 Boer), as if the
zodiac was equivalent to the equator and the hours were equinoctial, would explain the ascendant sign
given in these horoscopes.40
I have observed that in 8 cases out of the 21 (GH 345, Ba 18.1, Ba 18.2; POxy 4246, 4259, 4264, 4266a,
4270) the method of counting 15 degrees per hour does not yield the given ascendant, neither beginning
the count from the degree of the sun nor from the beginning of the sign.41 This is especially significant in
the three cases (GH 345, Ba 18.1, POxy 4266a) where the longitude of the sun is too close to the beginning
of the sign (between 5 and 9 degrees) to suppose that the astrologer had counted from the beginning of the
next sign. At least here we can be confident that this method was not applied. 42 43
End Beg.
Nr. Time Sun Asc. Whole sign 15-dg method
hour hour
Given ascendant sign coincides with the correct one
GH 316 5 day Lib 3 Sag Sag 6 Sco 24 Sag 5 Sag
GH 370 4 night Cap 19 Vir Vir 18 Vir 3 Vir 1 / Vir 29 Vir (Leo)
GH 376a 1 day Lib 20 Sco Sco 2 Lib 20 Lib 14 /Sco 12 Sco (Lib)
BA 18.242 8 night Sco 13 Sco Sco 16 Sco 2 Sco 5 Vir (Sco)
POxy 4257 3 night Gem 17 Cap Cap 23 Cap 10 Cap 5 Aqu (Cap)
POxy 4268 9 night Lib 1 Leo Leo 21 Leo 9 Leo 21 Leo
Given ascendant sign precedes the correct one
GH 345 8 day Can 5 Lib Sco 5 Lib 20 Lib 30 Sco
BA 18.1 6 night Aqu 9 Lib Sco 6 Lib 21 Lib 28 Sco
POxy 4247 5 night Sag 24 Leo Vir 9 Leo 24 Leo 18 Vir (Leo)
POxy 4248 4 day beg.43 Tau 27 Can Leo 1 Can 19 Can 10 Can (Gem)
POxy 4251 3 night Aqu 25 Vir Lib 7 Vir 23 Vir 14 Lib (Vir)
POxy 4259 6 day Sag 19 Aqu Pis 12 Aqu 25 Aqu 16 Pis
POxy 4264 4 day Ari 29 Gem Can 9 Gem 24 Gem 16 Gem

39 Cf. below, POxy 4266.1.

40 Equinoctial hours differ from the normally used seasonal in that, instead of being twelfth sections of day or night, are
th
24 sections of the whole day. Baccani hypothesizes that this method was often used in these horoscopes: cf. Ba, 79.
41 In the table I have given the two options in the case that the result would have been different (i.e. when the hour is odd).
In contrast with the whole-sign method, where the starting-point should naturally be the beginning of a sign, the 15-degree
method could have been measured from the rough degree of the sun. I have omitted the result using the beginning of the next
sign, which can be easily deduced from the other two values. In my reckoning, I have used the given hours as if they were
understood as equinoctial, assuming that the astrologers did not take the pains of converting the hours; otherwise the merit of
this approximate method – simplicity – would have been lost.
42 After the dot, I indicate the position of the concerned horoscope in papyri containing multiple horoscopes.
43 The Greek is περὶ ἄκρον (“on the edge”) which Jones translates “about the end”. The expression could however also
mean at the beginning (cf. LSJ, s.v. ἄκρον), and maybe the fact that the astrologer could have used the beginning of the same
sign and not of the next – which would have given Leo – points in that direction (but cf. below GH 282).
220 C. Tolsa

POxy 4266.1 10 day Sco 28 Aqu44 Ari 25 Ari 7 Pis 21 Ari

POxy 4266a 2 day Leo 9 Leo Vir 8 Leo 23 Leo 30 Vir
POxy 4269 2 day Aqu 29 Pis Ari 9 Pis 18 Pis 3 Pis
Given ascendant sign follows the correct one
GH 282 9 day beg.45 Vir 20 Aqu Cap 21 Cap 6 Cap 6 / Aqu 2 Aqu (Cap)
POxy 4246 7 night Cap 12 Sco Lib 27 Lib 11 Lib 16 / Sco 13 Lib (Sco)
POxy 4250 5 day Can 16 Lib Vir 30 Vir 15 Vir 16 / Lib 14 Lib (Vir)
POxy 4267 9 day Sco 17 Ari Pis 22 Pis 3 Pis 1 / Ari 9 Ari (Pis)
POxy 4270 7 night compl.46 Pis 29 Cap Sag 29 Sag 16 Sag 8 / Sag 30 Sag (Sco)
44 45 46

To summarize: it has been shown that astrologers used several strategies depending on the features of the
time reports given by their clients. First, we found two horoscopes probably using an astrological method
for the determination of the ascendant sign in the case of time reports lacking indication of the hour. In the
small sample of complete horoscopes using degrees, we have seen that in about half of the instances the end
or the beginning of the hour is indicated, and in the second case two degrees are added to the ascendant
corresponding to the stroke of the hour; in the other half, not showing such indication, it is likely that an
astrological method using the phase of the moon was applied to select a precise time within the given hour.
There is no evidence, then, of astrological methods that could result in an ascendant degree contradicting
the time record. As for the plain horoscopes, the great number of cases where the true ascendant at the end
of the indicated hour does not correspond to the indicated ascendant sign suggests that an approximate
method was normally used. Analysis shows that the one allotting 15 degrees to each sign in succession
cannot have been used in all cases, whereas all discrepancies can be explained by the use of a simple table
of rising times for whole signs. The resulting ascendant frequently falls one sign before the true ascendant,
working in practice as a randomizer selecting a time within the given hour. Therefore, it seems that the
given hour without further indication was generally not taken by astrologers to indicate the stroke of the
hour, but the whole interval.

Cristian Tolsa, Queen’s University, 505 Watson Hall, 49 Bader Lane, Kingston, Ontario K7L 3N6
cristian.tolsa@queensu.ca

44 This could be a mistake for Pisces.

45 In this case, it seems that the astrologer ignored the customer’s indication that the birth had taken place at the beginning
of the hour (ἀρχὴ θʹ).
46 The text specifies πλήραις (“completed”).