Vettius Valens’ Longitudes (1.

18),
Balbillus, and the Illusion of
Astrology’s Self-sufficiency
Cristian Tolsa

T HE PLANETARY LONGITUDES, that is, the positions of the

planets (including the sun and moon) on the path of the
sun—the ecliptic, or the zodiac, if conceived as divided
into twelve equal sections starting with the vernal equinox—
calculated for a client’s date of birth, are probably the most fun-
damental given with which astrologers operated (and operate) to
derive the effects of the planets on that person’s life. Thus, the
overwhelming majority of horoscope reports found in papyri,
the main direct witnesses of the spread of astrology in the
Graeco-Roman world apart from the theoretical astrological
manuals, consist solely of lists with these longitudes for each of
the planets, usually just giving the sign (e.g. Saturn in Aries, and
so on).
It might then come as a surprise that ancient astrological
manuals, both Greek and Roman, say nothing of the astro-
nomical methods by which such longitudes should be computed.
There is however one exception in the work of the second-
century astrologer Vettius Valens (1.18), whose mechanism and
underlying theory were reviewed in the Roman-era section of
O. Neugebauer’s monumental study of ancient astronomy.1 The
computation described by Valens is surprisingly simple, as can
be seen in the first method, for Saturn (1.18.6–9):

1 O. Neugebauer, A History of Ancient Mathematical Astronomy (Berlin/Heidel-

berg 1975: HAMA) 793–801. Valens: D. Pingree, Vetti Valentis Antiocheni Antho-
logiarum libri novem (Leipzig 1986).
—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
2019 Cristian Tolsa
398 VETTIUS VALENS’ PLANETARY LONGITUDES

τὸν µὲν οὖν Κρόνον οὕτως ψηφιστέον. τὰ ἀπὸ Αὐγούστου ἔτη

πλήρη ἀναλαβὼν, ἔκκρουε ὁσάκις δύνῃ ἀνὰ λʹ, τὰ δὲ περιλει-
πόµενα πολυπλασίασον ἐπὶ τὸν ιβʹ· καὶ ὁσάκις εἶ ἐκκεκρουκὼς
τριακοντάδας, ἑκάστου κύκλου ἀναλάµβανε εʹ καὶ ἀπὸ Θὼθ
ἑκάστου µηνὸς ἀνὰ αʹ, ἑκάστης δὲ ἡµέρας ἀνὰ λ<επτὰ> βʹ, καὶ
συγκεφαλαιώσας ἀπόλυε ἀπὸ Καρκίνου κατὰ τὸ ἑξῆς ἀνὰ λʹ· καὶ
ὅπου δ’ ἂν καταλήξῃ, ἐκεῖ ἔσται ὁ ἀστήρ.
Saturn is to be computed as follows: take the completed years
since Augustus and divide by 30 as far as possible. Multiply the
remainder by 12. Multiply the result of the division by 30 times
5. For each month from Thoth add 1°, and for each day 2ʹ.
Having totaled all this, count forward from Cancer, giving 30° to
each sign. The star will be where the count stops.
As Neugebauer shows, the results are at most approximations.
Far from using the well-attested Greek adaptation of the arith-
metical systems from the Babylonian theory,2 or the Babylonian
Goal-Year period of 59 years in which Saturn returns to roughly
the same longitude and stage in the synodic cycle—the cycle of
planetary phenomena in relation to the sun, such as stations, op-
position, conjunction—the method works with a sidereal period
of 30 years in which the planet returns, after one revolution, to
approximately the same point in relation to the fixed stars,
adding 5 additional degrees as a correction.3 The daily position
of the planet is then found by assuming constant velocity
throughout the whole 30-year period. If in 30 years the motion
is of 365 degrees, in one year it moves ca. 12 degrees, in one
month 1 degree, and in one day 2 minutes. This is obviously not

2 See A. Jones, “Studies in the Astronomy of the Roman Period III. Plan-
etary Epoch Tables,” Centaurus 40 (1998) 1–41.
3 For the use of Babylonian Goal-Year periods in Greek astronomy see

Ptol. Alm. 9.2, as well as the astronomical papyri edited and commented in A.
Jones, Astronomical Papyri from Oxyrhynchus (Philadelphia 1999) (where they are
used in conjunction with the arithmetical methods). For the Babylonian
theory of the Goal-Year periods see J. M. K. Gray and J. M. Steele, “Studies
in Babylonian Goal-Year Astronomy I: A Comparison between Planetary
Data in Goal-Year Texts, Almanacs, and Normal Star Almanacs,” Archive for
History of Exact Sciences 62 (2008) 553–600.

—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
 CRISTIAN TOLSA 399

the true longitude, since the planet does not appear to move at
constant speed throughout the zodiac, actually experiencing
retrogradation around opposition. However, because of
Saturn’s small change in velocity (it moves more slowly in
comparison with the other planets) the discrepancy is of just a
few degrees.4
The procedures for Jupiter and Mars are analogous to the one
for Saturn, but they attain lower degrees of accuracy, especially
in the case of Mars. For Jupiter, the 12-year sidereal period is
used (in which Jupiter completes approximately one revolution),
with an algorithm involving an intermediate step designed to
make the formula for the longitude appear very similar to that
of Saturn (HAMA 795), probably for the sake of memorization,
but at the cost of some extra inaccuracy. In the case of Mars, a
crude period of 2 years in which the planet supposedly completes
one exact revolution is used, giving a result involving an error of
more than one quadrant for the mean longitude (the longitude
of the epicycle’s center), but which in Valens’ example is suspi-
ciously made to agree with the true longitude by way of a couple
of apparent mistakes in the application of the rules (HAMA 796).
The approach for the inner planets is different. Venus com-
pletes 5 synodic periods (i.e. solar cycles) in 8 years, and the
method correspondingly divides 8 years into 5 roughly equal
periods of time (360 x 8/5=576º), each starting with a maximum
elongation as evening star (the phase when the inner planets are
at maximum distance from the sun) taking place at the beginning
of different zodiacal signs (HAMA 798). Neugebauer interprets
that the dates in Valens are garbled, but it seems possible that

4 In comparison with the epoch value here (i.e. the departing point of the
method), that is, Augustus 1 Thoth 1 = 0º Cancer, the mean longitude (i.e. of
the center of the epicycle) was 13º Gemini, and the true longitude 21º30ʹ
(HAMA 795), but these are tropical longitudes calculated from Ptolemy’s
methods (i.e. counting from the vernal point, whose slow motion through the
stars is called precession); the sidereal longitude normally used by ancient
astrologers including Valens would yield by the second century some 5 de-
grees higher, amounting in this case to approximately 27º Gemini, only 3
degrees from 0º Cancer.
—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
400 VETTIUS VALENS’ PLANETARY LONGITUDES

they were modified so that the longitudes were close to the first
degrees of the signs (see the computed longitudes in Neu-
gebauer’s table). Then, instead of using an auxiliary table with a
template to track the position of the planet along one synodic
period, there follows a procedure apparently designed ad hoc for
the given examples, since it subtracts 120 days to skip to another
phase of the planet (station as morning star) and then compute
the rest of days using a mean velocity valid for the period be-
tween this phase and the next maximum elongation. Nothing is
said for the case in which the date falls before that period of 120
days in the synodic period. The case of Mercury is perhaps even
more striking. Not only does the method assume an exact
number of synodic periods in a year (three), which with time
produces impossible results (viz. placing Mercury too far from
the sun), but again no template is used, just making the planet
advance with the mean velocity of the sun (one degree per day)
(HAMA 800). The procedure involves an addition of 162 days
that take the planet back to an initial position close to its phase
of maximum elongation as evening star, the same as Venus be-
fore: a possible conclusion is that the inventor of the method
drew inspiration from practice with auxiliary templates that gave
the advancement in longitude along the synodic cycle.
The question is, then, to what purpose were such methods
invented and used? It is obvious that Valens does not use them
for his horoscopes, which are all well computed using auxiliary
astronomical tables. Furthermore, as Neugebauer shows, Valens
in some cases manipulates the data exemplifying the methods,
so that they approximately agree with the real, computed longi-
tudes.5 He is then fully conscious that the methods were of little
practical use, and probably records them in part because they
appeared in the work of some predecessor. The epoch dates at
the beginning of the Augustan era mark a terminus post quem at
least for the version in Valens, and by that time arithmetical
schemes were already in circulation in the Graeco-Roman
world, as shown by the approximately correct longitudes of the

5 HAMA 796 (Mars), 799 (Venus).

—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
 CRISTIAN TOLSA 401

first horoscopes on record (cf. n.11 below for the horoscopes

from Balbillus’ manual). It could however be the case that astro-
nomical tables were not then as widespread as in the following
centuries. Maybe astrologers of that period would be more
prone to devising methods for the computations of planetary
longitudes that did not make use of auxiliary astronomical tables,
even if in the case of Mars and the inner planets such methods
were doomed to fail.
From the number of surviving papyri, it is possible to speculate
that an astrological boom in the first century A.D. in the Graeco-
Roman world, combined with a relatively low familiarity with
the technical needs of astrological practice, favored the inven-
tion of these methods. Of course, the numbers can be skewed
both by non-uniformity in the publication of papyri for different
periods and by non-uniform preservation contexts, hence A.
Jones’ caution when suggesting that the low quantity of astro-
nomical/astrological papyri in the first centuries B.C. and A.D. is
caused by the lower general survival rates of papyri.6 However,
it is likely that the practice of astrology was not yet widespread
in the first century B.C., given that Hypsicles adapted the Baby-
lonian ascensional times for the zodiacal signs to the latitude of
Alexandria around the middle of the second century B.C. The
first century A.D. is by far the most fruitful in the number of
astrological authorities,7 which corresponds quite well with a
surge in astrological interest among the imperial elite (Tiberius,
Nero)8 and with the high relative frequency of astronomical and
astronomical papyri charted below. The fact that the high num-
ber of such papyri in the third century is not paired with that of
astrological authorities is perhaps due to the use of astrological
manuals from the preceding centuries.

6 Jones, Astronomical Papyri from Oxyrhynchus 6–7.

7 Dorotheus, Manilius, Anubio, Thrasyllus, Balbillus, Teucer.
8 For a classic analysis of the political and intellectual context of this period

in relation to astrology see F. Cramer, Astrology in Roman Law and Politics (Phila-
delphia 1954).

—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
402 VETTIUS VALENS’ PLANETARY LONGITUDES

Century Number of Number of Relative Relative

astronomical documentary frequency frequency
papyri + papyri9 with 30-year
horoscopes factor10
I B.C. 7 6370 0.11% 0.07%
I A.D. 30 10951 0.27% 0.21%
II A.D. 46 26522 0.17% 0.16%
III A.D. 74 23427 0.32% 0.28%
IV A.D. 29 16556 0.18% 0.26%
By sheer coincidence, among the first surviving Graeco-
Roman horoscopes are two from an excerpt of Ti. Claudius
Balbillus, active towards the middle of the first century A.D., and
who, as I will claim, was plausibly linked to the invention of these
methods.11 The dates of the horoscopes (72 B.C. and 43 B.C.) are
exceptionally early for an astrologer active under Claudius and
Nero. Neugebauer and van Hoesen argue, probably rightly, that
they were likely taken from the archives of his father Thrasyllus,
the famous astrologer friend of Tiberius, but this does not ex-
plain why Balbillus did not use horoscopes of his own as was
normally the case with practicing astrologers. My hypothesis is
that he was aware of the enormous deviations of his rules of
thumb for some of the planets in comparison with the true
longitudes, and, not willing to manipulate the data in the way
that Valens did, picked horoscopes from before the time of
Augustus, so that those methods could not be directly applied.
We are fortunate enough to have an early summary of Bal-
billus’ astrological work, which ends with the following remark
9 Source: papyri.info, retrieved 2018/12/20.
10 In this column I take into account the fact that astronomical/astrological
papyri are dated by the astronomical data from several decades before (the
client’s birth date), redistributing the number of horoscopes so that in average
30% in each century are passed to the next, i.e. I assume that charts were on
average cast 30 years after the date of birth. Thus, e.g. for I A.D. 30% of the
30 go to II A.D., while 30% of the 7 from I B.C. are added here, giving 23.1.
11 Analysis and translation of these horoscopes in O. Neugebauer and H.

B. van Hoesen, Greek Horoscopes (Philadelphia 1987) 76–78. Cf. M. Gansten,

“Balbillus and the Method of aphesis,” GRBS 52 (2012) 587–602. On the
figures of Balbillus and his father Thrasyllus in relation to Roman politics see
again Cramer, Astrology in Roman Law and Politics.
—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
 CRISTIAN TOLSA 403

(CCAG VIII.3, 103–4):

παραδίδωσι δὲ καὶ διαφόρους µεθόδους εὑρέσεως ὡροσκόπου
καὶ µὴν καὶ καταλήψεως τῶν ἀστέρων ἀναλύσεως ἄνευ.
And he gives different methods for finding the ascendant, as well
as for apprehending the [longitudes of the] planets, without retro-
gression.
There are several reasons to believe that the latter is a reference
to the methods copied by Valens. First, no other ancient astro-
logical work that I know of teaches the reader how to use
astronomical tables to find the longitudes of the planets. It was
understood that this was the domain of astronomy. The same is
true for the methods to find the ascendant: there is the normal
astronomical procedure, employing the astronomical table of
ascensions, but as the text refers to “different methods” (διαφό-
ρους µεθόδους), it almost certainly means the so-called natural
methods, through which astrologers adjusted the ascending
point. Ptolemy justifies the use of these methods, arguing that
sundials were notoriously inaccurate,12 a circumstance related to
the fact that the time of birth was customarily reported to the
precision of whole seasonal hours (which theoretically give much
room to place the ascendant, namely one-half zodiacal sign on
average).13 Likewise, the summary likely refers to methods for
finding the longitude of the planets such as the above, which do
not use auxiliary astronomical tables, and since we do not know
of other such methods, there is a fair probability that these were
the ones copied by Valens. Also, the fact that the epoch dates
are Augustan fits Balbillus’ chronology excellently. If that were
not enough, there is still the intriguing phrase ἀναλύσεως ἄνευ,
“without ἀνάλυσις,” which qualifies the methods in the sum-
mary. The concept of ἀνάλυσις covered a variety of meanings,
all of them technical and almost exclusively applied to (1)

12 Ptol. Tetr. 3.3.1–2 (ed. W. Hübner, Claudi Ptolemaei Opera III.1 [Leipzig

1998]). R. Hannah, Time in Antiquity (London/New York 2009) 106, estimates

an average error of 15 minutes.
13 On this matter see C. Tolsa, “Time of Birth and Ascendant in the Pa-

pyrus Horoscopes,” ZPE 204 (2017) 209–220.

—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
404 VETTIUS VALENS’ PLANETARY LONGITUDES

mathematical methods for the solution of problems or theorems

going backwards to already known facts in reversible steps, or (2)
the reduction of philosophical propositions to already-estab-
lished first principles. Neither of them seems to be directly
applicable to methods for the determination of the planetary
longitudes, but the general sense of going backwards could have
evolved here (and perhaps in other, non-attested instances) to
signify the planetary retrogradation, which is as a norm called
ἀφαίρεσις (“subtraction”) on account of the decreasing longi-
tude of the retrogradating planet. As demonstrated above,
among other inaccuracies, our methods do not allow for
retrogradations, perhaps the most striking of their deviations.
Supporting the possibility of this meaning for ἀνάλυσις I can
adduce the evidence from second- and third-century writers on
theology who use the term as a substitute for ἀφαίρεσις to refer
to the definition method for the divine entity that proceeds by
subtraction of negative attributes, the so-called negative
theology practiced by Middle Platonic thinkers and Christian
theologians.14
Secondly, among the natural methods for determining the
ascendant called “gnomon” by Valens, one is attributed to Thra-
syllus, who, as Valens suggests, only recorded it privately (Vett.
Val. 9.11.1–4):
καὶ ἑτέραν ἀγωγὴν ὑπό τινων αἰνιγµατωδῶς ἀναγεγραµµένην
ὑποτάξω … δύσκολον µὲν οὖν καὶ ἐργῶδες ἀλλοτρίας δόξας
ἐλέγχειν, καὶ ταῦτα µηδὲ διὰ γεγραµµένων βιβλίων µηδὲ διὰ
λόγων ἐνεργητικῶν παρειληφότα, καθάπερ ὁ Πετόσιρις τῷ βασι-
λεῖ περὶ πολλῶν µυστικῶς ἐκτίθεται … πολλὰς οὖν αἱρέσεις
ἐκτίθεται ἐξεπίτηδες διά τε τοὺς µεµυσταγωγηµένους καὶ ἀπαι-
δεύτους, ὧν τὴν δύναµιν οἵ γε νοῦν ἔχοντες εὐκατάληπτον ἕξου-
σι, καὶ ἃς µὲν ἰδίας ἃς δὲ λεληθότως ἀναγεγραµµένας.
I will also append a method recorded mysteriously by some
astrologers … It is difficult and laborious to prove a doctrine from
elsewhere, especially if it has not been received through written

14 Clem. Strom. 5.11.71.2–3; Celsus ap. Origen C.Cels. 7.42.9–11. Cf. D.

Jugrin, “The Way of ἀνάλυσις: Clement of Alexandria and the Platonic Tra-
dition,” Studia Philosophiae Christianae 52.2 (2016) 71–94, at 73 and 77.
—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
 CRISTIAN TOLSA 405

books or powerful dialogues, in the way that Petosiris mystically

exposes many theories to the King … [The compiler] pur-
posefully exposes many methods for both the initiates and the
ignorant, whose utility will be grasped by the intelligent, and some
of them are written privately, others secretly.
Then he explains the method and finally reveals that Thrasyllus
used it. It turns out that we also have a summary of Thrasyllus’
work, which, in contrast to that of Balbillus, appears to be ex-
haustive; and in fact, as Valens implies, Thrasyllus does not seem
to have written on the topic of the natural determination of the
horoscope.15 If he recorded it privately, then a plausible scenario
is that his son Balbillus knew the method and copied it in his
work, in the section devoted to such doctrines in the summary,
citing his father as the inventor. Hence, Valens may well have
used Balbillus’ work.
Third, the method of the gnomon ascribed to Thrasyllus by
Valens bears an interesting resemblance to the methods for the
longitudes of the planets, in that its most distinctive feature is the
use of divisions and their remainders. Basically, it instructs to
measure the elongation (the distance along the zodiac) in degrees

15 CCAG VIII.3 99–101. It is the common view that Thrasyllus’ work was

named Pinax, from this witness alone—the title of the excerpt reads συγ-
κεφαλαίωσις τοῦ πρὸς Ἱεροκλέα Θρασύλλου πίνακος, and a very similar
statement closes it. I suspect however that πίναξ, meaning “astrological
board,” was also a generic name for an astrological treatise. The evidence is
in the fourth-century compiler Hephaestion, who on only two occasions
seems to give the title of an astrological work. Since most astrological writers
had written only one book, their names were sufficient and titles were entirely
forgotten if they ever existed; Hephaestion’s own treatise bears no title. One
of his apparent titles corresponds to the Synagogai of Protagoras of Nicaea
(Apotel. 3.47.52), which suspiciously has the same title that Hephaestion
(3.pr.1) once applies to his own work. The other is Critodemus’ Pinax (Apotel.
2.10.42), but, as we know from Valens (3.9.3, 9.1.5), Critodemus’ work was
entitled Horasis, a much more specific and thus credible title. The author of
the summary must have felt compelled to choose a descriptive term (yet not
necessarily the exact title) because he wanted to include the name of the
dedicatee. The same would be true of Balbillus’ ἀστρολογούµενα (CCAG
VIII.3 103–104).

—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
406 VETTIUS VALENS’ PLANETARY LONGITUDES

from the sun to the moon, and to compare this with the re-
mainder of dividing by 360 the multiplication of the length of
the day by the hour of birth. The distance from the sun to the
moon is also used in a method by Balbillus’ contemporary
Anubion to find the ascendant naturally (ap. Heph. Apotel.
2.2.11–18); the fact that this number appears independently in
two roughly contemporary sources could imply that it was de-
rived from an older source.16 The second number—the re-
mainder of dividing by 360 the multiplication of the length of
the day by the hour of birth—is much weirder. The length of the
current day runs from 170 to 210 time-degrees in the latitude of
Alexandria, and the multiplication by the hour of birth can
reach a very large number, which is then reduced modulo 360.
Methods of this kind only survive recorded in Valens.
Fourth, in the chapters in Valens’ text preceding the expo-
sition of the rules of thumb for the longitudes of the planets, we
find a series of methods for the natural determination of the
ascendant containing variations of the one by Thrasyllus (1.4),
as well as a chapter on the calculation of the longitude of the
moon (1.17), both of which could have been taken from the last
part of Balbillus’ astrological work. The method for the moon,
which gives good approximations, uses the fact that 19 years
contain an almost exact number of lunations (i.e. synodic periods
16 Valens specifies that the distance must be taken in time-degrees, that is,
in equatorial degrees, which recalls the standard procedure of “primary
directions” for measuring the client’s lifespan from the distance between the
starting-point (ἀφέτης) and the destroyer (ἀναιρέτης), actually the first doc-
trine recorded in Balbillus’ summary. Astrologers chose these two points
along the ecliptic and then computed the equatorial degrees in which this
stretch of the zodiac rose above the horizon, using a table of ascensions that
divided the whole rising time of a sign equally among all ecliptic degrees.
Thus, if Scorpio rises in 35 time-degrees, any degree of Scorpio was assumed
to rise in 35/30 time-degrees (this is, of course, a simplification based on the
mathematical average; in the astronomical reality even the single degrees
have slightly different rising times). Then these degrees were counted as years
of life. The procedure was refined with Ptolemy’s table of ascensions in his
Handy Tables, which uses spherical trigonometry, but this was not widespread
until the middle of the third century.

—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
 CRISTIAN TOLSA 407

of the moon), so that from the date and longitude of an initial

conjunction, one deduces the day since the last conjunction and
the longitude with respect to the sun (applying the rule of
12º/day) by counting the years, months, and days within the
present 19-year cycle. However, as Neugebauer explains, in-
stead of simply reducing the latter number modulo 29.5 (the days
of a complete lunation), Valens employs an apparently un-
necessary complication by dividing the number of completed
years within the present cycle by 3, and instructing the reader to
look at the remainder. I will not repeat Neugebauer’s exposition,
but it is important to keep in mind that if the remainder is 1, one
is then asked to use the number 10, and if it is 2, the number 20;
and nothing if it divides evenly.17
If we now go back to the summary of Balbillus’ treatise, we see
that he recorded a doctrine called ἐξάλµατα (“leaps”) (CCAG
VIII.3 104):
µεθοδεύει δὲ καὶ περὶ ἐξαλµάτων λέγων οὕτως· τὰς ἀναφορὰς
ἑκάστου ζῳδίου δίπλωσον καὶ ποίησον παρὰ τὸν πέµπτον. καὶ
ἐὰν µὲν καταλείπῃ α´ ἡµέρα, θὲς ἀντὶ µιᾶς ιʹ· ἐὰν δὲ βʹ, κʹ· οἷον
<εἰ> Διδύµοι ἀνατέλλουσι κηʹ, δὶς δὲ κηʹ, νϛʹ· ταῦτα παρὰ τῶν εʹ
µέριζε ἑνδεκάκις πέντε, γίνεται νεʹ, λοιπὸν αʹ, ἥ ἐστιν τῶν ιʹ· εἰ
δὲ βʹ ὑπελείποντο, κ’ ἂν ἦσαν καὶ ἑξῆς ὁµοίως.
Then he deals with the leaps, saying the following: double the
ascending degrees of the sign and divide by 5. If the remainder is
1 day, instead of 1 use 10; if it is 2, 20. Thus if Gemini ascends in
28, the double of 28 is 56; divide this by 5, 11 times 5 makes 55,
the remainder is 1, this is the 10. If the remainder is 2, it would
be 20, and similarly with the rest.
From the summary, we cannot know what this was used for. All
we know is that for each sign, one picks the rising times (the
equatorial degrees it takes to rise), doubles them, divides the
resulting number by 5, and uses the result of 10 times the re-
mainder. We immediately see the striking similarity with the
counterintuitive reckoning of the moon’s longitude in the above
method. It is very specific that the wording in the final step, the
17 Vett. Val. 1.17.1: ἐὰν µὲν γὰρ περισσεύῃ αʹ, πρόσθες ιʹ τῷ ἀριθµῷ· ἐὰν

δὲ βʹ, κʹ· ἐὰν δὲ γʹ, µηδέν.

—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
408 VETTIUS VALENS’ PLANETARY LONGITUDES

multiplication by 10 of the reminder, is the same, for it would

have been more natural to say “multiply the remainder by 10”
than “if it is 1, take 10,” and so on.
We can go a bit further to understand the doctrine presented
here by Balbillus. The word ἐξάλµατα recurs in an astrological
context only in Paulus’ Elem. (92.10–22 Boer) and its commen-
tary attributed to Heliodorus (131.1–10 Boer). Paulus explains a
doctrine he names “circumambulation” (περίπατος) of the sun,
moon, and ascendant, which is analogous to the computation of
the years of life. Using the sun (supposedly if it is a nocturnal
nativity one uses the moon), he measures the time-degrees from
its position (in the direction of the signs) to all hostile points on
the zodiac (in his example a point square with Mars) to find the
dates of life’s crises. The only difference with the usual calcula-
tion of the length of life is that here we take any hostile point,
whereas, in the calculation of the length of life, the so-called
destroyer (ἀναιρέτης), which naturally marks the ultimate crisis,
needs to meet more conditions.
Paulus adds that attaining the limit of a new zodiacal sign,
what he calls a leap, also indicates a crisis. So, if the sun is at 15º
Scorpio, the 15º remaining until Sagittarius are counted as 35
(the time-degrees in which Scorpio ascends) divided by 30,
multiplied by 15 = 17.5, meaning that in 17 years 6 months there
will be a crisis. Now the commentator explains another method
for the leaps, “resulting in the same” (τὸ αὐτὸ πάλιν ἀνάγουσα),
which consists in multiplying the rising time by 2, then by 6, and
counting the result as days. These are then divided by 30, and
the resulting number of months is assigned to each degree of the
sign. It is clear that Heliodorus here modifies Balbillus’ leaps to
make them compatible with the standard method of counting
time-degrees, as used by Paulus. This is revealed by the fact that
he first doubles the ascensional times (like Balbillus) instead of
multiplying directly by 12, and by the contradiction of naming
the result of this operation “days.” If the rising time of Scorpio
is 35, that means 35 years, which multiplied by 12 should yield
420 months, not days. This is because Balbillus’ procedure gives
directly the days corresponding to each degree.
—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
 CRISTIAN TOLSA 409

It thus seems that Balbillus was keen on playing with multi-

plications, divisions, and residues to build periodical distribu-
tions of numbers, making of this a kind of trademark, unless this
was also typical of other astrologers who have left no traces in
the extant record. By doubling the ascensional times of the signs,
then dividing by 5 and multiplying the remainder by 10, he
attained a qualitatively different result than if he had not first
doubled:18
Signs With doubling Without doubling
Aries, Pisces 50 50
Taurus, Aquarius 30 40
Gemini, Capricorn 10 30
Cancer, Sagittarius 40 20
Leo, Scorpio 20 10
Virgo, Libra 50 50
From the numbers in the chart, we should probably interpret
that for Balbillus the term “leaps” referred to the leaps between
the quantities in the sequence of the signs, not the leap to the
end of the signs in the computation of the crises as in Paulus.
Paulus thus substituted Balbillus’ weird assignment of times for
the usual one, transferring the concept of leap to the assignment
of a crisis at the end of each sign. The end of a sign then probably
did not mark a crisis for Balbillus, who, if I am right here, was
very conscious of the numerical distribution resulting from his
procedure, naming the doctrine accordingly.
In any case, a similar kind of creativity is attested in the in-
vention of the label “gnomon” for each of the numbers that need
to be compared in order to astrologically determine the ascend-
ing degree. Since the ascending point resulting from the table of
rising times is interpreted to be correct in the case when the
numbers coincide, it is clear that the name alludes to the pointer

18 I am using the latitude of Alexandria, and assume that when the division

was exact the remainder was understood to be 5, not 0. In the method for the
longitude of the moon, the remainder of division by 3 of the number of years
in the present cycle of 19 years can likewise be 3 (Vett. Val. 1.17.1–3).
—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
410 VETTIUS VALENS’ PLANETARY LONGITUDES

of a sundial, which actually gave the hour of birth. Some sundials

like the Tower of the Winds in Athens, but also the more typical,
portable kind, were formed of multiple dials and pointers, so a
usual procedure to check the accuracy of the time is likely to
have consisted in comparing more than one dial. If they agreed,
there was a good chance that the hour was correct. The astro-
logical gnomon seems to be based on the same principle.
It is worth emphasizing that these two labels (leaps and gno-
mon) share something quite specific. Whereas most astrological
doctrines bear more general names—houses, terms (ὅρια), meth-
od of aphesis, etc.—leaps and gnomons refer to concrete, intrinsic
features of the numerical method employed. This adds plausi-
bility to the thesis that we are dealing with astrological creations
from the same specific milieu (Thrasyllus/Balbillus), the same
milieu in which the similarly-working algorithms for the planets
and the moon were likely devised.
Conclusions
Returning to our initial problem, it is worth noting that the
attempt to find simple numerical algorithms for determining the
planetary longitudes represents an instance of two general trends
in Hellenistic astrology. One is the idea of simplification, the
other a certain kind of mystification. The doctrine of the terms
is a case in point, where we find both things. Each sign was di-
vided into unequal sections, each ruled by a different heavenly
body. Hellenistic-Egyptian astrologers inherited from Baby-
lonian sources a system for this which cannot be derived from a
small set of rules, that is, it can only be given term by term in a
table,19 in contrast to systems attested only in Greek sources,
generated from different kinds of periodical rotations. It is telling
that Ptolemy (Tetr. 1.21) struggles with the so-called Egyptian
system, the one derived from Babylonian texts, and presents
another system, based on a simple rotation of the ruling planets,
which he confusingly calls Chaldean but which is most probably

19 A. Jones and J. Steele, “A New Discovery of a Component of Greek

Astrology in Babylonian Tablets: The Terms,” ISAW Papers 1 (2011), http://

doi.org/2333.1/k98sf96r.
—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
 CRISTIAN TOLSA 411

Hellenistic. Finally, he proposes a “rational” modification of the

Egyptian system purportedly encountered in a damaged old
manuscript. This one is probably a hoax, since no actual division
of the signs can be attained with the rules he proposes, and the
table in the manuscripts has been incorporated from later com-
mentaries.20 Consequently, Ptolemy’s preferred terms were not
used until later commentators established a concrete system
which only approximately corresponded with Ptolemy’s rules.21
We can also qualify some of the present methods for the
discovery of the planetary longitudes as a hoax, since they do not
even remotely give the correct longitudes. If my analysis is
correct, Balbillus provides us with easy, “rational” methods
based on his usual astrological procedures, which supposedly
serve to determine longitudes which every trained astrologer
knows must be discovered with complex astronomical tables
using Babylonian numerical methods. The method for the moon
seems unnecessarily complicated, but the astrologer probably
sought to find an algorithm that looked similar to the procedure
for the leaps. Likewise, the rule for Jupiter contains an apparent
complication in fact aimed at giving it the same appearance as
the one for Saturn. The fact that the computation of the moon’s
longitude, also using these periods, is effective, suggests that
Balbillus could have tried to derive the astronomical data naively
with these procedures. After all, astrologers were in general not
very good astronomers. Seeing that these methods approxi-
mately worked for Saturn and Jupiter, he would have silenced
the complete lack of agreement for the rest of the planets. The
same seems to be the case with Ptolemy and his terms. A real
effort on his part is seen in the confection of the rules for his
desired system, but not reaching a satisfactory result apparently
did not stop him from giving the directions for the derivation of
his terms. The inclusion of a hoax could then form part of the

20 C. Tolsa, “The Table of Ptolemy’s Terms,” Philologus 162 (2018) 247–

264.
21 S. Heilen, “Ptolemy’s Doctrine of Terms and its Reception,” in A. Jones

(ed.), Ptolemy in Perspective. Use and Criticism of his Work from Antiquity to the Nine-
teenth Century (New York 2010) 45–94.
—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
412 VETTIUS VALENS’ PLANETARY LONGITUDES

generic characteristics of astrological writing: perhaps it was a

game readers were used to and even expected and enjoyed.
Valens can be useful for understanding this context. He often
describes astrological texts (both his own and others’) as having
two kinds of readers in mind, beginners and initiates. Speaking
of Critodemus’ Horasis, he claims that its opening, with a
reference to the astrologer’s voyage (supposedly in a search for
knowledge), sounds fantastical and marvellous to learners,22 and
also notes that some of Critodemus’ readers “have discarded his
empty words and tracked down the relevant chapters with great
toil.”23 Valens also frequently mentions that his predecessors
spoke in riddles (µυστικῶς), at one point acknowledging that he
is also writing in this guise in some parts of his text in order to
generate interest among his students.24 One facet of the writing-
style aimed at beginners may have been to pretend that astrology
was self-sufficient with respect to astronomy. Ptolemy is the only
astrological author who assumes right away that astrology is
secondary with respect to mathematical astronomy (Tetr. 1.1),
but precisely because he was first and foremost a great astron-
omer he can be the exception that confirms the rule. We have
already seen that the fact that the ascending point could not be
attained with precision led to the invention of astrological meth-
ods for its adjustment.
These widespread techniques—to judge from the many meth-
ods preserved in the treatises, but also from some documentary
horoscopes25—could have convinced some that astrology was
not subordinated to astronomy, or even was more precise. For
example, Valens warns against astrologers who criticize the
makers of astronomical tables “either from jealousy or crooked-

22 Vett. Val. 3.9.3: ἐντυχέτω δέ τις τῇ λεγοµένῃ ὁράσει Κριτοδήµου, πῶς

µὲν τὴν ἀρχὴν εὐφαντασίωτον ἔχει καὶ τὰ λοιπὰ τετερατολογηµένα πρὸς τοὺς
ἀµαθεῖς.
23 Vett. Val. 3.9.5: οἱ µὲν τὰς µαταιολογίας παραπεµψάµενοι καὶ δι-

εξιχνεύσαντες τὰ δοκοῦντα κεφάλαια µετὰ παντὸς πόνου.

24 Vett. Val. 2.3.1 (Nechepsos); 2.41.2 (Petosiris); 7.6.2 (Nechepsos and

Petosiris); 9.4.3 (Zoroaster). See 3.10.19, speaking of himself.

25 Cf. Tolsa, ZPE 204 (2017) 209–220.

—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
 CRISTIAN TOLSA 413

ness” (διὰ φθόνον ἢ διὰ τὸ σκολιὸν τῆς εἰσόδου), demanding

that his readers “stick to exactness” (τὸ ἀκριβὲς προσέχειν); but,
at the same time, he seems just thereafter to fall into the same
mistake when he mentions his attempt at correcting small errors
in Apollinarius’ tables by looking at the natural effects of the stars
(6.4.8–9). In another passage, he remarks on the variations
among astronomical tables of his time and the lack of consensus
over the length of the years among astronomers, which, he says,
led him to attempt the confection of his own tables for the sun
and the moon (9.12.8–9):
ἐπειράθην µὲν οὖν καὶ αὐτὸς κανόνα συµπῆξαι Ἡλίου τε καὶ Σε-
λήνης πρὸς τὰς ἐκλείψεις· ἐπεὶ δέ µε ὁ χρόνος περιέκλειε τὸ τέλος
ἐπάγων…
I tried myself to calculate an astronomical table of the sun and the
moon using eclipses, but since the time was too limited to bring
this to completion…
Then he provides the names of the astronomers responsible for
the tables he used. It would be naive to think that he is speaking
frankly here: Valens clearly had not the astronomical acumen
needed to produce good astronomical tables. He rather wants to
impress the reader by posing as an even better astronomer than
the best astronomers of his time. For this reason I am inclined to
believe that the methods for the planetary longitudes are not just
a pia fraus intended not to deter beginners from the study of
astrology and designed as a substitute for using proper astro-
nomical tables which would come later in the training process.
My claim is that they rather aim at presenting astrology as
superior to astronomy in every aspect.
The circumstance that Valens offers Balbillus’ methods,
which, as we have seen, he knows do not work, should then not
be surprising, for it would be part of the same game. It should
not bother us that he also recommends using astronomical tables
and sticking to astronomical exactness, since contradiction is not
a problem if one operates at two levels; Valens actually repre-
sents one of the most extreme cases of tolerance for contradiction
in antiquity. For example, he harshly criticizes Critodemus’
voyage-description, big tables, and astrological oaths as arrogant
—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414
414 VETTIUS VALENS’ PLANETARY LONGITUDES

tricks to attract readers, but he ends up incorporating his own

variations of these three devices in his text.26 His reason to do so:
as he has expressed in his criticisms, he thinks that these devices
impress beginners. And what an impression would it make to
reveal that planetary longitudes could be obtained with those
simple rules, which only astrologers know!27

May, 2019 Universität Osnabrück

ctolsa@gmail.com

26 The description of the voyage, the oaths, and the tables are criticized at

3.9.3–5; the voyage and the tables again at 9.1.5–7. Valens mentions his own
arrival in Egypt, echoing Critodemus’ words, at 4.11.4; he exacts an oath at
4.11.12 and 7.1.1–3; and he draws two big tables at the end of Book 8, the
first of which we know (from the title of 3.6) is from Critodemus.
27 I wish to thank the two GRBS reviewers for their many insightful com-

ments contributing to the improvement of this article.

—————
Greek, Roman, and Byzantine Studies 59 (2019) 397–414