Athanase Papadopoulos

MENELAUS’ SPHERICS IN GREEK AND ARABIC
MATHEMATICS AND BEYOND
ATHANASE PAPADOPOULOS
Abstract. We present the history of Menelaus’ Spherics (2nd c. AD), ex-

plaining its importance and quoting in detail some of the significant propo-
sitions. We include this work in the general context of Greek and Arabic
mathematics, highlighting the far-reaching connections with modern math-
ematical ideas.
This paper will appear as a chapter in the Handbook of the History and
Philosophy of Mathematical Practice (ed. Bharath Sriraman), , Springer,
to appear in 2024.
Keywords. Menelaus of Alexandria - Spherics - Spherical geometry -

Spherical trigonometry - Greek mathematics - Arabic mathematic - Theo-
dosius of Tripoli - Autolycus of Pitane - Pappus of Alexandria
AMS codes. 01A20 - 01A30 - 5303 - 53C70
1. Introduction
Menelaus’ Spherics (2nd c. AD) constitute one of the most remarkable
achievements of Ancient Greek mathematics, along with Euclid’s Elements,
Apollonius Conics, Diophantus’ Arithmetica, Archimedes’ treatises on ge-
ometry and statics, Theodosius’ Spherics, Pappus’ Collection, to name only
some of the most important Greek mathematical works. Among these trea-
tises, Menelaus’ Spherics is certainly the least understood by historians and
the least familiar to mathematicians of the modern period. As a matter of
fact, we know that Fermat was nourished by the work of Diophantus, that
Christiaan Huygens was deeply immersed in the writings of Archimedes and
Apollonius, that Euler was an assiduous reader of Diophantus and Pappus
and that he was well acquainted with Theodosius’ Spherics, not to mention
of course Euclid’s Elements, a work which all mathematicians of the past
thoroughly studied. But very few of the modern mathematicians mention
Date: February 21, 2023.

1
2 ATHANASE PAPADOPOULOS
Menelaus’ Spherics,1 a fact that might seem surprising, especially since a rel-
atively large number of medieval Arab mathematicians were well acquainted
with this work.
I can see two reasons for which Menelaus’ Spherics remains unknown.
First, there are problems inherent to the available versions: no Greek man-
uscript has survived, and the ones we possess, in Arabic or translated from
the Arabic, present intricate differences in form and wording. Another rea-
son is that this work is difficult to access from the mathematical point of
view. Indeed, the existing manuscripts do not contain proofs of several of
the difficult propositions. I will say more on this below.
Menelaus’ Spherics opened up a new research field. Saying this, we are
not talking about the subject of the geometry of the sphere which, as a
general topic, already existed (we mentioned Theodosius’ Spherics, but the
sphere was studied before Theodosius, by Autolycus, Euclid, Archimedes
and others), but of the field known today as the “intrinsic geometry of sur-
faces”, which can be described, precisely like Menelaus investigated the ge-
ometry of the sphere, as the study of the properties of curvilinear triangles
on a surface which is not the Euclidean plane. Indeed, the vast majority
of Menelaus’ propositions are concerned with such properties. Menelaus’
major achievement is that he highlighted for the first time the notion of
spherical triangle, and he developed the bases for its study.
The Arabs included this work into their body of knowledge and they
understood its novelty. Let us quote right away Ah.mad ibn Sa‘d al-Harawı̄
(10th c.) from the very beginning of his edition of Menelaus’ treatise (English
translation in [31, p. 500]):
The way that Menelaus followed in that book is marvelous; nobody
preceded him and nothing had been known to his predecessors on
that subject. Did he foresee that this kind of geometrical science
is unique in itself, that it has laws suitable to it and lemmas that
lead to the goal which are different from those that one is willing
to demonstrate by using straight lines and the determination of
planes by their intersections?
The sentence about“straight lines and the determination of planes by their
intersections” refers to the fact that Menelaus avoided as much as possible
the use of Euclidean solid geometry which is the ambient geometry of the
1An exception is Marin Mersenne, who was fond of old mathematical texts and who was
aware of this work. Indeed, Mersenne published in [22] a Latin edition of the third book of
the Spherics, together with works of Autolycus, Archimedes, Apollonius, Theodosius, Dio-
phantus and Pappus. But one century later, Anders Lexell, who was one of Euler’s closest
collaborators, and who was a specialist of spherical geometry, writes in his memoir [21]:
“From the time where the Spherical elements of Theodosius were handed in written form,
it is very difficult to find, in the treatises of the geometers, other elements concerning fur-
ther refinements of the theory of the figures drawn on the surface of the sphere, than those
that are usually exposed in the Elements of Spherical Trigonometry and that are aimed at
solving spherical triangles.” It is not completely clear to what work with the title Elements
of Spherical Trigonometry Lexell refers (this may be Euler’s Principes de la trigonométrie
sphérique tirés de la méthode des plus petits et des plus grands [13]), but it is obvious that
he was not aware of Menelaus’ work, written two centuries after that of Theodosius, nor
of the works of Nas.ı̄r al-Dı̄n al-T . ūsı̄ (d. in 1274) and the other Arab mathematicians of
the 11th-13th century who worked extensively on spherical trigonometry.
MENELAUS’ SPHERICS IN GREEK AND ARABIC MATHEMATICS 3
sphere; the expression “straight lines and planes” refers to Euclidean lines
and planes. Let us also quote Menelaus himself, from the oldest fragment
known of the Spherics, which we also edited and translated into English in
[31, p. 504]:
I have invented a kind of proof which is excellent and admirable.
Several things relative to surface <properties> on the spherical
surfaces appeared to me; I don’t think they were offered to anybody
else. And I ordered the lemmas and the proofs in an order so
that it will be easy for those who are fond of science to rise up
and attain universal and honorable sciences. In this way, with
the help of particular propositions that occur to them from the
experiences they acquire from that art, they will become skillful
and outstanding in the universal propositions, given the multitude
of universal propositions that can occur to them, and that become
for them proofs and indications for the inaccessible and the difficult
directions in the principles of Spherics, their role and what ensues
from them.
The field that Menelaus opened in his Spherics can be especially appre-
ciated in the context of non-Euclidean geometry, a domain of mathematics
that became a particularly active research field in the last quarter of the
twentieth century.
In the pages that follow, I will review this work in its context, discussing
at the same time the mathematics it contains, the history of the texts that
we possess, and the impact of this work on the Arab mathematicians of the
period 9th-13th c., who are certainly, among the mathematicians who read
this work, those who have most understood its importance.
The content of the rest of this chapter is the following:
In §2, I briefly review works on the geometry of the sphere by two prede-
cessors of Menelaus, namely, Theodosius of Tripoli and Autolycus of Pitane,
pointing out relations between these works and Euclidean geometry, making
connections with the work of Menelaus, with a later work by al-H . asan Ibn
al-Haytham and hinting to some relations with the development of modern
analytic spherical geometry.
Section 3 contains some preliminary general facts about Menelaus’ Spher-
ics, including its division and its main characteristics.
Section 4 is the heart of this chapter. I discuss in it some significant
propositions of Menelaus’ Spherics and at the same time I give an overview
of the major themes that are addressed in this work. I believe that before
commenting on the history of this book, one has to understand the subject it
treats, and for this, a knowledge of at least some of the important statements
and methods of proof is necessary.
I have divided this book, as in my edition with Rashed of al-Harawı̄’s
rectification [31], into seven major themes. The reader interested in a more
complete discussion of the mathematical content of this work is referred to
Part II (p. 125-395) of our edition.
It may be noticed that some of the statements of Menelaus’ propositions
are analogues of propositions in Euclid’s Elements (e.g. the triangle in-
equality, equality and inequality criteria for triangles, etc.), but one must
be aware of the fact that this is only an apparent similarity in the sense
that the proofs of the spherical propositions are very different from those
of their Euclidean analogues. This is due to the fact that the properties of
the underlying surfaces are very different: On the one hand, we have the
Euclidean plane, in which two lines may intersect or not, and where any two
lines starting at the same point diverge, and on the other hand, we have the
sphere, in which any two lines intersect and where two lines starting at a
point start by diverging and then converge to the same point. Furthermore,
in the Euclidean plane, the angle sum in a triangle is constant and equal
to two right angles, whereas on the sphere, the angle sum in a triangle is
non-constant and is always greater than two right angles. There are many
other substantial differences between the two geometries.
In §5, I discuss the influence of Menelaus’ Spherics on Pappus. This is a
rare occasion where the influence of this work on a later author from Greek
Antiquity is visible.
Sections 6 and 7 constitute an important part of this chapter; they con-
cern the influence of Menelaus’ Spherics on Arabic mathematics of the period
that starts at the 9th century and that ends at the 15th, where a consid-
erable amount of work was done in translating this treatise and in trying
to complete the proofs of the difficult propositions that are contained in it.
This effort led to a development of important mathematical theories by the
Arabs, especially in spherical geometry.
In §8, I discuss the connection of the Spherics with astronomy, a relation
which, in my opinion, has been exaggerated by modern authors.
In the following, when I quote mathematical propositions, I have some-
times paraphrased the original text to make it more easily readable by a
modern mathematician. The translations close to the original together with
the original text of Menelaus’ Spherics are contained in the volume [31].
Acknowledgements This work obviously owes a lot to Roshdi Rashed

with whom I collaborated for the edition [31], and I take this opportunity to
express my admiration for all the work he has done for the history of Greek
and Arabic mathematics.
2. Two predecessors: Theodosius and Autolycus

Before talking about Menelaus’ Spherics, I will make a short review of the
works of two forerunners, Theodosius of Tripoli (1st c. AD) and Autolycus of
Pitane (4th c. BC). I will start with Theodosius, the immediate predecessor
of Menelaus, mentioning his in influence on the latter, and then, I will
consider Autolycus, on whose work was based that of Theodosius.
Theodosius’ Spherics was written in the first century BC, that is, two
centuries before Menelaus. The two works, which carry the same title, are
concerned with the properties of the sphere. Theodosius’ treatise was obvi-
ously not foreign to Menelaus. In fact, Menelaus in a more than one sense
has placed himself in the continuity of Theodosius: It suffices to recall that
at the beginning of Book II al-Harawı̄’s rectification (a version of this work
to which we shall often refer), Menelaus declares that his goal is to improve
and generalize some propositions of Theodosius [31, p. 684]. Furthermore,
Menelaus, at some places of his Spherics, refers either explicitly or at least
implicitly to Theodosius’ work, and uses some of his propositions. We have

commented at length on this fact in our edition [31] of al-Harawı̄’s rectifica-
tion, and we have also highlighted there the major differences between the
two works.
The best modern-language edition of Theodosius’ Spherics is still Ver
Eecke’s [33]. It very close to the text from which it is translated, it repro-
duces faithfully the mathematical content of the work, and it is accompanied
by a solid mathematical commentary.2 Let me start with some general re-
marks on this work.
Theodosius, in his Spherics, makes a distinction between a theorem and a
construction. Such a distinction, which has already been made explicit and
commented upon by Proclus in his Commentary to the first book of Euclid’s
Elements, is highlighted by Heiberg and by Ver Eecke.3 The latter, after the
number of each proposition, writes whether it is a theorem or a construction.
An example of a construction in Theodosius’ Spherics is Proposition 2 of
Book I, whose statement is: To find the center of a given sphere. This is an
analogue of Proposition I of Book III of Euclid’s Elements whose statement
is: To find the center of a given circle. An important remark that fits here is
that in Menelaus’ Spherics, which is the main topic of the present chapter,
all the propositions are meant to be constructive.
Theodosius’ Spherics is divided into three books. None of the statements
of the propositions in the first two books is close to any statement of a
proposition of Menelaus’ Spherics. In this respect, the two treatise are very
different. If we want to summarize in very few words the difference between
them, we can say that the first one is concerned with circles drawn on the
surface of the sphere whereas the second one is concerned with triangles
drawn on the surface of the sphere. This is especially true for the first two
books of Theodosius.
Let us highlight some statements of Theodosius’ Spherics.
Book I starts with definitions: that of the sphere (by which is meant the
solid sphere), its surface, its center and the notion of diameter. After the
definition of the latter, it is said that it is a rotation axis of the sphere, a
sentence which Ver Eecke considers as a later addition ([33, p. 1, Footnote
3]). Indeed, the property that a diameter of the sphere is a rotation axis
is not used in the treatise. It is used in Autolycus’ work which we review
below.
Theodosius defines the pole of a circle on the surface of the sphere as
a point on this surface such that all the Euclidean segments starting at
this point and ending at a point on the circle are equal. In particular, no
attempt is made to define the pole intrinsically on the surface of the sphere
(for instance, as a center with respect to lengths of arcs of great circles on
that surface).
Theodosius then gives the definition of “two planes having the same incli-
nation.” His definition is the same as Euclid’s. We recall that the definition
2There are editions and commentaries done by authors with a poor knowledge in math-
ematics, and I prefer not to dwell upon this.
3Ver Eecke’s edition is based in part on Heiberg’s, see the comments in [33] p. lii of
the Introduction.
of the angle made by two planes is contained in Euclid’s Book XI (Definition

6), and that the definition of two pairs of planes having the same inclination
is the content of the next definition of the same book. We also note that
unlike in Euclid’s Elements, besides this relative notion of inclination, there
is no absolute notion of angle made by two planes in Theodosius’ Spherics,
neither in the definitions nor in statements or in the proofs. In his propo-
sitions and in his proofs, Theodosius only talks about perpendicular great
circles and oblique great circles. The notion of angle between two lines is
only used for the Euclidean lines in the solid sphere, and never for two arcs
of circles on the surface of the sphere. These constitute fundamental differ-
ences between the Spherics of Theodosius and Menelaus. Indeed, Menelaus
uses the dihedral angle between two planes to define the angle made by two
circles on the sphere, and hence, to define the notion of angle at the vertex
of a spherical triangle, and he applies this notion in an essential way in every
proposition and in every proof, with very few exceptions.
After the definitions, the first book of Theodosius’ Spherics, in Ver Eeckes’
edition, consists of 23 propositions.4 These propositions concern basic prop-
erties of the sphere and its intersections with planes. For instance, Proposi-
tion I.1 says that the intersection of the sphere with a plane is a circle. We
already mentioned Proposition I.2, which is a construction problem: To find
the center of a given sphere. A great circle is the intersection of the sphere
with a plane passing through the origin. Proposition I.11 says that on the
sphere, two great circles cut each other in two equal parts. Proposition I.12
is a converse and is less trivial: it says that if two circles on the sphere cut
each other into halves, then these two circles are great circles. Proposition
I.13 says that if, on a sphere, a great circle cuts another circle with right
angles, then it cuts it into two equal parts and passes by its poles. Propo-
sition I.16 says that the distance from a great circle to a pole is equal to
the side of a regular inscribed quadrilateral (i.e., a square). Proposition I.17
is a converse. Proposition I.18 is a construction problem: To construct a
segment equal to the diameter of a given circle on the sphere. Proposition
I.21 is also a construction problem: To find the pole of a given circle on the
sphere.
From the mathematical point of view, all this is Euclidean geometry; there
is no spherical geometry at all.
Let us pass now to Book II.
This book, in Ver Eecke’s edition, also consists of 23 propositions. Most
of them deal with properties of small circles. The book starts with the
definition of tangent circles: two circles are tangent to each other if the line
that is the intersection of the two planes that contain them is tangent to
both of them. Proposition II. 6 says that if a great circle is tangent to a
small circle, then it is also tangent to another small circle equal to the first
small circle. I have reproduced the figure of this proposition from an Arabic
manuscript, Ms. 3464, Topkapi Saray, Ahmet III (Istanbul). This is the
drawing on the top in Figure 1. I would like to take this opportunity to
make a comparison with the figure of the same proposition, as redrawn in
4The attribution of the last two propositions to Theodosius, according to Ver Eecke, is
doubtful, see [33, Footnotes p. 29 and 30].
Figure 1. From Theodosius’ Spherics. On the top, the fig-

ure for Proposition II.6. Ms. 3464, fol. 30, Topkapi Saray,
Ahmet III.
the plate of the figures of Clavius’ English edition of the Spherics [8], which
I reproduced here in Figure 2. In this plate, this is No. 33.5 The figure in
the ancient Arabic manuscript shows a sense of abstraction which, in my
opinion, is absent from the realistic figures in the more recent editions of
Clavius, Ver Eecke, etc.
Let me mention two other propositions from the second book of Theodo-
sius’ Spherics, which are both construction problems. Proposition II.14 says
the following:
5In Clavius’ edition, the total number of propositions contained in each book is not the
same as in Ver Eecke’s edition.
Figure 2. A plate of figures for Book II of Theodosius’

Spherics, Clavius’ edition [8].
Given a circle on the sphere which is not a great circle and given a point
on it, to describe a great circle tangent to the given circle at the given point.
Proposition II.15 is more involved:
Given on the sphere a circle which is not a great circle and given a point of
the sphere situated in the region between this circle and a parallel equal circle,
to describe a great circle tangent to the first small circles and passing through
the given point. These propositions have a taste of Apollonian constructions
of circles tangent to each other in the plane. The figure that comes with the
last construction, in the Istanbul 3464 Ms., is reproduced in Figure 3 below.
I have included it here to suggest the complication of the construction. This
is also Figure 42 in Clavius’ version (Figure 2 here).
The content of Book II also belongs to the realm of 3-dimensional Eu-
clidean geometry.
Book III of Theodosius’ Spherics is concerned with important monotonic-
ity properties that we discuss at length in [31]. Such properties constitute
Figure 3. From Theodosius’ Spherics, Proposition II.15.

Ms. 3464, fol. 30, Topkapi Saray, Ahmet III.
the culmination of both Spherics, in terms of degree of difficulty. Let me

explain by an example what I mean by a monotonicity property. I take as
example Proposition III.9 in Theodosius. The corresponding figure, from
the manuscript mentioned above, is reproduced here in Figure 4. The 3
sub-figures correspond to three cases in the proof, and we do no not need to
worry about this now. In each case, the action takes place in a hemisphere,
bounded by the exterior great circle represented, and the proposition con-
cerns two arcs of great circles intersecting each other with an acute angle,
one which we shall call the equator, and another one, the oblique arc (re-
ferring naturally to Figure 4). In each of the three sub-figures, the point on
the upper left is a pole of the equator. On the oblique arc we take three
points and we consider the two consecutive segments they bound. (There
are several segments represented in the figure, and they are all used in the
proof, but the statement involves only two of them.) We project these two
Figure 4. From Theodosius’ Spherics, Proposition III.10.

Ms. 3464, fol. 50, Topkapi Saray, Ahmet III.
segments perpendicularly (that is, using lines emanating from the pole),
onto the equator. The proposition says that the ratio of the segment which
is closer to the acute angle to its projection is greater than the ratio of the
other segment to its projection. The word “monotonicity” that we used to
describe the theme of this proposition refers to such an inequality. More pre-
cisely, this word refers to the fact that if we start, instead of two consecutive
segments on the oblique line, with a sequence of more than two consecutive
segments on this line, then the sequence of ratios of one segment to its pro-
jection are all different and they vary monotonically: the more a segment is
close to the acute angle of the triangle, the more the corresponding ratio is
greater. We note that in the analogous Euclidean situation where we project
perpendicularly segments on a Euclidean straight line onto another straight
line, the conclusion of the statement is that the ratios are all equal.
Despite the fact that his methods were criticized by Menelaus who ob-
tained more general results and who proved them using intrinsic methods,
Theodosius remains, to the best of our knowledge, the first mathematician

who highlighted such monotonicity properties which are proper to spherical
geometry. We mentioned above that from the purely mathematical point
of view, in Books I and II of Theodosius’ Spherics all the statements are
Euclidean. On the contrary, this proposition and other similar ones in Book
III brings us to the heart of spherical geometry.
Between the Ancient Greek period and the modern one, monotonicity
properties properties in Theodosius’ Book II have been featured by Ibn al-
Haytham, the preeminent 10th-century Arab mathematician, see [28], in
particular p. 49-105 of the French edition. Roshdi Rashed, the author of
this edition of Ibn al-Haytham, considers (rightly) that the developments
by Ibn al-Haytham that were based on this proposition of Theodosius and
similar ones constitute an early chapter in infinitesimal geometry. Note by
the way the related multiplicity of lines in Figure 4, which suggest a passage
to the limit.
Let me mention here that several modern treatises on hyperbolic geometry
start with propositions similar to Theodosius’ III.9 that we just mentioned,
and that such propositions are used in the foundations of the infinitesimal
theory of non-Euclidean theory. For instante, in the doctoral dissertation of
Louis Gérard, Sur la géométrie noneuclidienne [16], defended in 1892 and
written under the direction of Poincaré, a monotonicity property similar
to III.9 is used for establishing a functional equation which gives directly
the trigonometric formulae of non-Euclidean geometry. In fact, the devel-
opment in Gérard’s thesis is done in the context of Lobachevsky geometry,
but the same arguments, with inequalities in the reverse direction, give the
trigonometric formulae for spherical geometry. I mention this here in order
to suggest the power of these inequalities. Later authors have also used the
same approach in the analytic development of non-Euclidean geometry, see
e.g. Barbarin’s book [2], one of the early French treatises on non-Euclidean
geometry, in which the author makes at the same time a comparison be-
tween the spherical and the hyperbolic cases, starting precisely with these
monotonicity properties.
From Proposition III.9 and its figure, it is also clear that spherical trian-
gles without the name are present in Theodosius’ Book III. Ver Eecke, on
p. x-xi of the Introduction to his edition of Theodosius’ Spherics, rightly
states that the latter was surely aware of some of the properties of spherical
triangles, but that he did not include this object in his work because he did
not succeed in identifying their important properties.
Book III of Theodosius’ Spherics contains 14 propositions, some of which
are much more involved than the one we mentioned. Menelaus, in his Spher-
ics gave improved and stronger versions of some of these propositions. We
shall review this in the next section, but let me mention here, in this connec-
tion, the beginning of Book II of Menelaus’ Spherics, in al-Harawı̄’s edition.
We read the following on the critique of Menelaus towards Theodosius, re-
garding the method used to prove these propositions [31, p. 684]:
Ah.mad ibn Abı̄ Sa‘d al-Harawı̄ said that Menelaus overcame the
difficulties of this science, something which was not within the
reach of anybody else. Despite the fact that he mastered it, and
albeit the majesty of what he was undertaking, he bypassed nu-

merous lemmas which are needed by someone who examines this
book and does not have the rank of Menelaus
We can see him criticizing Theodosius in his book on the Spher-
ics; he sees that the method that he followed is not satisfactory,
since it comprises complications and the drawing of numerous
<straight> lines, and that Theodosius did not take into account,
by this method, the properties of the figures which take place in
the sphere, that is, the states of the angles generated by the inter-
section of circles.
[I swear] on my life that everything that Theodosius proved in
that book, Menelaus proved it in an easier way, and he intended
that the proof is by the direct method.6
Theodosius’ work was influenced by that of Autolycus, in particular, his
two treatises On the moving sphere and On risings and settings written
about 300 hundred years before.
Autolycus lived in the 4th century BC, about the same time as Euclid. His
treatise On the moving sphere contains 12 propositions and is considered as
the oldest complete Greek treatise that reached us. The expression “moving
sphere” in the title denotes the Celestial sphere which, according to the
theory that was in use at that time, rotates uniformly around an axis that
crosses the Earth which sits at the center of the universe and which passes
through the poles.
From the mathematical point of view, the twelve proposition of Autolycus
are easy. I will just limit myself to three propositions from Autolycus’ work,
my aim being just to give an idea of some of the background of the two
Spherics, especially the one of Theodosius. The reference I use is [1].
Proposition 1 says the following:
If a sphere rotates with a uniform velocity around an axis, all the points
situated on the surface of the sphere and that are not on the axis describe
parallel circles that have the same poles as the sphere and that are perpen-
dicular to the axis.
Ptoposition 6 says:
If on the (moving) sphere the fixed great circle (that is, the horizon)
is oblique to the axis, it is tangent to two equal parallel circles: the one,
situated on the side of the visible pole, will always be visible, and the other
one, situated on the side of the invisible pole, will always be invisible.
One may compare this proposition to Theodosius’ Proposition II.8 which
we quoted above.
Proposition 12 of Autolycus says:
If on the (moving) sphere a fixed circle always cuts into two equal parts
an arbitrary moving circle of the sphere and if none of the two is neither
perpendicular to the axis nor passes by the pole of the sphere, then the two
circles are great circles.
The reader may notice the relation between the last proposition and
Proposition I.12 of Theodosius’ Spherics which we quoted above.
6“By the direct method” means without using a proof by contradiction.

Some propositions of Autolycus’ Moving sphere are used in Theodosius’

astronomical treatises, On habitations and On the days and nights which we
mention in §8.
We pass now to Menelaus’ work.
3. Menelaus’ Spherics: Preliminary remarks

Talking about Menelaus’ Spherics, one must necessarily mention the Arab
commentators of this work, since this work reached us only trough a certain
number of Arabic versions; all of them containing more or less substantial
differences among themselves (and consequently with respect to the original
Greek text).
We start by talking about the division of this work.
There is no general agreement on the division of Menelaus’ Spherics. Nas.ı̄r
al-Dı̄n al-T.ūsı̄, in the preface of his edition which we quote in [30, p. 13],
writes the following:
This treatise consists of three books in some copies, and of two
in some others. Concerning the three books, in most cases, the
first one contains thirty-nine propositions, the last one twenty-
five propositions, and the middle one, in many copies, contains
twenty-four propositions, and in the copy of Ibn ‘Irāq, twenty-
one propositions. And in some of them, the first one contains
sixty-one propositions, the second one nineteen propositions, and
the third one twelve propositions. Regarding the copies with two
books, the first one contains sixty-one propositions and the last one
thirty propositions. And there are differences in some propositions,
because some of them made two propositions a single one and
conversely. But in general, the total number of propositions in the
treatise is between eighty-five and ninety-one. I referred in the
margin to some of the books and the number of propositions they
contain, in red and in black, and sometimes I did it in the text.
Ibn Abı̄ Jarrāda, in a manuscript we quote in [31, p. 17] (see also §6
below), mentions a division into four books.
We have discussed at length the divisions of the various versions of the
Spherics that survive in Chapter I of our edition of al-Harawı̄’s rectification
[31], which is divided into two books. At the beginning of Book II, a sentence
explains this last division [31, p. 684]:
As we have shown the lemmas which we needed (i.e., in the first
61 propositions), let us return now to what Theodosius wanted to
show; we prove it using a universal assertion, without admitting in
it the impossible.7 We shall then show his error and rectify what
is spoiled.
We conclude from this that Menelaus suggests that his goal in Book II is
to improve some propositions of Theodosius. A rapid inspection of the two
treatises shows that the propositions in question are those of Book III of
Theodosius’ Spherics. Indeed, Book II of Menelaus’ Spherics contains a
certain number of propositions that are concerned with the comparison of
projections of segments of a great circle onto another such segment, in the
spirit of Proposition III.9 of Theodosius’ Spherics which we recalled above.
7This means by a direct method, without using proofs by contradiction.
Although al-Harawı̄’s rectification of the Spherics is divided into two

books, the numbering of the proposition is continuous, from 1 to 91.8 In
what follows, it will be practical for us to use this numbering.
Now about the content.
The most important object of study in Menelanus’ Spherics is the spher-
ical triangle.9 This is a figure made by three arcs of great circles called its
sides, each two sides intersecting at a point called a vertex of the triangle.
A side of a triangle is always assumed to be less than a semicircle. In this
way, a side is the unique shortest arc between its vertices. This justifies the
fact that in what follows, we shall sometimes call segment an arc of great
circle which is smaller than a semi-circle, and angle an angle made by two
segments at an intersection point.
Menelaus avoided as much as possible proofs that use solid Euclidean
geometry. This is the point of view of intrinsic geometry of surfaces which
we mentioned in the introduction. Furthermore, Menelaus always sought for
new proofs. In the few cases where a proof similar to that of an analogous
proposition in Euclid’s Elements is possible in the spherical setting, he either
does not give any proof at all (like in Proposition 6 which says that if we
draw from the two vertices of the base of a triangle two arcs of great circles
that meet within the triangle, then the sum of these arcs is smaller than
the sum of the two legs of the triangle), or he gives a different proof (like in
Proposition 8 which says that if two triangles have two pairs of equal sides
such that the angle contained by one pair belonging to one of these triangles
is greater than the angle contained by the corresponding pair in the other
triangle, then the base of the triangle that has a greater angle is greater
than the base of the other triangle).
It is conceivable that Menelaus did not give a proof of the famous Sector
Figure (Proposition 66) because he did not have a proof which does not
use the ambient Euclidean geometry. It is also possible that Ptolemy, who
uses extensively this proposition in his Almagest, gave a proof of it (see [17,
p. 50–55]) because of the absence of proof in Menelaus’ treatise. We shall
mention several times this proposition in what follows.
Furthermore, a large number of arguments of Menelaus rely heavily on
polarity properties on the sphere that do not hold in the Euclidean place.
Thus, in most of the cases, the proofs of Menelaus’ propositions of spher-
ical geometry cannot be used for their Euclidean analogues. We shall give
explicit examples in the next section.
4. The propositions
There is more than one way of grouping the propositions of Menelaus’
Spherics into topics: such a grouping depends on the importance that we
attach to each particular theme and on the relations we can make between
the various topics. I have collected the propositions of Menelaus’ Spherics
into seven groups, corresponding to seven themes, which I review now.
8Regarding the number of propositions, one should note that there are statements for
which Menelaus gives two different proofs, and each proof is counted as a new proposition
(this is the case, for example, of Propositions 8 and 9). Thus, in some sense, there are less
than 91 propositions.
9Two Arabic expressions I ÊJÓ (threefold) and ¨C @ éKCK ð X É¾ (“trilateral figure”)
are used by the Arabic translators of Menelaus for the Greek term τ ρίπλυρoν (“trilateral
”). Let us recall that Euclid, in the Elements (Definition I.3), defines trilateral figures as
those contained by three straight lines.
Figure 5. Figure for Proposition 41
Group 1: Construction of angles; Propositions 1, 41-44. I start with

Proposition 1: Given an arbitrary angle, and given an arbitrary point on a
given line, we can construct an angle equal to the given angle having its
vertex at the given point and a side on the given line.
This proposition is of paramount importance, and it is abundantly used
in the proofs of the rest of the propositions of the Spherics. Its proof, unlike
the vast majority of the other propositions, uses the solid geometry of the
ambient space. More precisely, it relies on constructions from Theodosius’
Spherics and a criterion from Euclid’s Elements for the equality of two angles
seen from the center of a circle.
A remark is necessary here. We mentioned that most of Menelaus’ proofs
are intrinsic, that is, they use notions defined on the surface of the sphere
(lines, angles, circles, etc.) and not from the ambient space. However, this
cannot be true for all propositions. The reason is that Menelaus did not
start with axioms for spherical geometry, since he thought of the sphere
as a surface embedded in Euclidean 3-space, therefore the axioms of solid
Euclidean geometry are sufficient for the study of such an object. Thus,
some of the propositions, and in particular the first one, necessarily rely on
Euclidean geometry. Establishing an autonomous set of axioms for spherical
geometry is a modern idea.10
Let us now pass to another proposition from the first group, namely,
Proposition 41. The general idea is to construct, in a spherical triangle, a
segment (that is, an arc of great circle) from a vertex to its opposite side,
making with this side an angle equal to one of the angles of the triangle at
the extremity of this side. The construction is not possible in the general
case, and I will explain why, but let me first state the proposition under
Menelaus’ hypotheses:
Given a triangle ABC such that AB + BC is smaller than a semi-circle
and such that the angle A is obtuse, and given a point D on BC, to construct
a line from D that cuts the side AB with an angle DIC [ equal to the angle
A of this is triangle.
10
An early formulation of a set of axioms for spherical geometry is in [18].
Figure 6. Proposition 41 fails with the assumption that AB

and BC are equal to a quarter of a circle
Figure 7. Proposition 41 fails when AB and AC are greater

than a quarter of a circle
The proof is contained in [31, p. 628-630]. I have reproduced here

Menelaus’ figure (Figure 5), just to suggest the fact that the construction is
involved. This proposition is used in later propositions of the Spherics, and
in particular in the last propositions.
Let me make a few comments on Proposition 41.
The Euclidien analogue of this proposition does not need any particular
hypothesis: one takes the parallel to AB passing through D; it makes with
AC an angle equal to the angle at A of the triangle ABC. But in the
spherical case, parallel lines do not exist, and the problem is much more
complicated. Without specific assumptions, the desired line does not always
exist: See Figure 6 for the particular case where each of AB and BC is a
quarter of a circle. In this case, a great circle from D cannot make with AC
a right angle, because D is not a pole of AC (since the pole is at B). In the
case where we allow AB + BC to be greater than a semi-cercle, the situation
is not better: Take for example the case where the angle A is acute and C
right, with AB and BC both greater than a quarter of a circle and let D be
a pole of AC (Figure 7). Then every line from D to the side AC makes a
right angle with it, therefore this angle cannot be equal to the angle at A.
In Proposition 42, the same construction is done under the hypothesis
that the angles at B and C are both acute. This proposition is used later
(e.g. in the proof of Proposition 45). I have reproduced Menelaus’ figure
in Figure 8 here. I have also reproduced (Figure 9) the page of an Arabic
Figure 8. Proposition 42
manuscript containing the three propositions, 42, 43, 44 on the same subject.
Notice the concision of the writing (the three propositions fit in one page).
Group 2: Comparaison between triangles; Propositions 4, 8, 9,
13-19 and 20-23. This group consists of congruence and inequality criteria
for spherical triangles. More precisely, there are two sorts of theorems in
this group: (1) congruence theorems, i.e. theorems saying that if we have
two triangles such that some elements (side lengths or angles) of one are
equal to the corresponding elements of the second, then the two triangles
are congruent; (2) non-congruence theorems, i.e., theorems saying that if we
have two triangles such that some elements (side lengths or angles) of one
are greater than the corresponding elements of the second, then some other
element of the first is greater (or smaller) than the corresponding element
of the second.
An example of a congruence theorem is Proposition 19, which has no
analogue in Euclidean geometry. It says that if two spherical triangles have
equal angles, then they are congruent.
We note incidentally that Menelaus never makes statements such as “the
triangles are congruent”, or “the triangles are equal”, or equivalent expres-
sions. Instead, he always gives precise statements, saying what side in the
first triangle is equal to what side in the second triangle and what angle in
the first one is equal to what angle in the second one.
Examples of non-congruence theorems are Propositions 8 and 9, which are
in fact a single theorem concerning two triangles having two pairwise equal
pairs of sides. They say that if in one of these triangles, the angle contained
by two equal sides (the legs) is greater than the angle in the second triangle
contained by a pair of legs equal to the pairs of legs in the first triangle,
then the first triangle has a greater base than the second. This proposition
also holds in Euclidean geometry (Euclid’s Proposition I.24).
Group 3: Geometry of triangles; Propositions 2, 3, 5-7, 10-12 and
24-26. These propositions concern the geometry of individual triangles. For
instance, Propositions 2 and 3 concern isosceles triangles. They say that if
a triangle has two equal sides (the legs), then these legs subtend with the
base equal angles, and conversely, if a triangle has two equal angles at the
Figure 9. Propositions 42, 43 and 44. Ms. 3464, fol. 88,

Topkapi Saray, Ahmet III.
base, then the two legs subtending them are equal. These are analogues
of Propositions I.5 et I.6 of Euclid’s Elements. The proofs though are very
different in the spherical case; they are based on the duality between lines
and their poles, and they also use some propositions in solid geometry from
Theodosius’ Spherics. Proposition 5 is the triangle inequality for spherical
triangles. Let us dwell on Proposition 11, which is used in a crucial way at
several places in the Spherics:
Let ABC be a spherical triangle. Consider the exterior angle BCD.
\ Then
we have the following three equivalences:
(1) BCD
\=A b if and only if AB + BC is a semi-circle;
(2) BCD
\>A b if and only if AB + BC is smaller than a semi-cercle;
(3) BCD
\<A b if and only if AB + BC is greater than a semi-cercle.
Let us prove the first statement. We produce the sides AB and AC until
we obtain a lune. One of the vertices of this lune is A, and we may set the
point D to be second vertex (Figure 10). If BCD \ = A, b then BCD
\ = D. b In
the triangle CBD, since C b = D,
b we have BC = BD (this is the congruence
criterion of Proposition 3). Since AB + BD is a semi-cercle, AB + BC is
also a semi-cercle.
The converse is obtained by reversing the arguments and using the con-
gruence criterion of Proposition 2. The proofs of (2) and (3) are of the same
type, based on the notion of lune and using Proposition 7 (inequality criteria
for triangles).
Let me remark that if one proves (1) and (2), then (3) follows immedi-
ately by a reductio ad absurdum, but Menelaus does not allow this kind of
reasoning.11 This is why a proof of (3) is needed.
Proposition 12 of the Spherics is of paramount importance in spherical
geometry; it says that the angle sum in any spherical triangle is greater than
two right angles. It is a direct consequence of Proposition 11. We reproduce
this proof, because it is also a noteworthy illustration of the intrinsic methods
of Menelaus.
To prove Proposition 12, let ABC be a spherical triangle and let us con-
sider the exterior angle BCD.
\ It suffices to prove that BCD \<A b + B.
b
We suppose without loss of generality that BCD \ > A (otherwise the
b
conclusion we seek is obviously true). Then, we construct an angle DCE \=
Ab < DCB
\ with E on the extension of the segment AB, as in Figure 11.
Now Proposition 11 applied to the exterior angle DCE \ of the triangle AEC
implies that AE + EC is a semicircle. Thus, BE + EC is smaller than a
semicircle. Applying again Proposition 11, this time to the triangle BCE,
gives CBA
\ > BCE. \ Adding A b to both sides gives BCD
\<A b + B,
b which is
the desired result.
11Regarding this fact, let us quote Ibn Abı̄ Jarrāda: “I said that Menelaus presented
numerous lemmas for the proof of this proposition; but al-Harawı̄ had shown some of them
by a reductio ad absurdum; and we said that this does not belong to Menelaus’ method. I
showed all what was presented in the course of the proof without a reductio ad absurdum.”
(see the exact reference in [31] p. 41).
Group 4: Division of triangles; Propositions 27-40. This group con-

sists of a set of propositions that concern inequalities obtained by dividing
a spherical triangle by a segment (that is, an arc of great circle) connecting
a vertex to the opposite side. The arc may be a side bisector, or an angle
bisector, and there are also other kinds of division of a triangle, for instance,
by an arc of gret circle joining the midpoints of two sides. I will report more
precisely on some of these propositions.
I start with a proposition which I find particularly interesting, namely,
Proposition 27. It says the following:
In a spherical triangle ABC, let D and E be the midpoints of AB and
BC, and DE an arc of great circle joining them. Then DE > AC/2.
The importance of this kind of proposition in modern metric geometry
cannot be overemphasized. It is equivalent to a concavity property for dis-
tances in spherical geometry, and more generally, in a space of positive
curvature. Herbert Busemann took the conclusion of Proposition 27 as a
definition of a metric of positive curvature, see [5, 6, 7].12
The proof of this proposition is a beautiful example of the intrinsic meth-
ods of Menelaus:
We produce ED until a point G such that DG = DE and we produce
AG and CB until they meet at a point H (Figure 12).
12Busemann was not aware of the fact that the property was stated explicitly in a
theorem of Menelaus. I highlighted this fact in [24].
Since DG = DE, AD = DB and EDB \ = ADG, \ the two triangles EDB

and ADG are congruent (This is one of the congruence theorems, Proposi-
tion 10 of the Spherics). Thus, we have EB = AG and EBA \ = DAG. \
From Proposition 11 on exterior angles which we recalled above, applied
the triangle HBA and its exterior angle B,b AH + HB is a semi-circle. Thus,
AH + HE is greater than a semi-circle. Now we join A and E by a segment
of great circle. Applying again Proposition 11 to the triangle HEA, we get
CEA
[ < EAG. [
Now the two triangles CEA and EAG have a common edge EA, and
we have CE = AG, and CEA [ < EAG. [ Therefore, AC < EG (this is
Proposition 8 of the Spherics). Since DE = 12 GE, then DE > 21 AC and
the proposition is proved.
Let us note that to perform the construction in the proof of Proposition
27, the sides of the triangle ABC must be small enough (for instance, we do
not want the point H to fall inside the triangle ABC, which could happen if
the segment AH is very long). We also note that in the definition of positive
curvature one can assume the triangles small (the definition is local).
It is also important to know that there is an analogue of this proposi-
tion in the setting of hyperbolic geometry (the other classical non-Euclidean
geometry), with the inequality in the conclusion reversed.
A more precise form of Proposition 27 can be proved using spherical
trigonometry and a formula of Euler giving the area of a spherical triangle.
I have included this proof in the first appendix to this chapter.
Among the other propositions in the same group, Propositions 28 and 29
concern again a triangle ABC with a segment DE joining the midpoints of
AB and CB, but here, instead of comparing the segment DE to the side
AC, one compares angles. More precisly, the propositions give the inequality
BDE
\ < BAC \ under various conditions on the angles of the triangle ABC.
Note that in the Euclidean case, we have the equality BDE\ = BAC. \
A recent paper provides analogues of propositions 28 and 29 of Menelaus’
Spherics in the setting of hyperbolic geometry, see [32]. I am mentioning
this to point out that such statements are still actual.
Group 5: Monotonicity; Propositions 45-62, 64, 65, 81-91. This

group contains 32 propositions that express monotonicity properties. We
already explained, in §2, what we mean by such properties and we talked
about their importance. At the basis of such a property, we generally have
a spherical triangle and projections of segments contained in one side onto
another side. Such projections may use arcs of great circles making equal
angles with the base side on which we project, like in Proposition III.9 we
quoted from Theodosius Spherics, but there are other kinds of projections:
using arcs of great circles perpendicular to the base, using small circles
parallel to a side, etc. All these monotonicity results provide inequalities
that show fundamental differences between spherical geometry and the two
other classical geometries (Euclidean and hyperbolic). Furthermore, like
for Proposition 41 which we recalled above, various kinds of hypotheses
on the shapes of the triangles (angles or side lengths) are necessary in the
statements of these propositions.
There is a progression in the results of Group 5 in terms of the sophistica-

tion of the statements and difficulty of the proofs. Some of these propositions
are improvements of propositions of Theodosius on the same topic. I will
mention two of these propositions, limiting myself to the statements.
Proposition 82 says the following:
Let ABC be a triangle with a right angle at A and acute angle at B,
and with BC at most equal to a quarter of a circle. Consider on BC two
arcs BD and EG and take the perpendiculars DH, EI, GK to the base AC
(Figure 13). Then,
if BD = EG then AH > IK and AB − DH < EI − GK;
if AB − DH = EI − GK, then BD > EG;
if BD + (AB − DH) = EG + (EI − GK), then BD < EG;
if BD − (AB − DH) = EG − (EI − GK), then BD < EG.
Furthermore, in every case, we have
crd 2BD crd 2(AB − DH)
> .
crd 2EG crd 2(EI − GK)
Proposition 89, in the same group, says the following:
Let AB and BC be two arcs of great circles. Consider two points D and
E on AB and take the perpendiculars DC, EH on BC (Figure 14). Then,
crd2CH
is equal to the ratio of the product of the diameter of the sphere by
crd2DE
the diameter of the (small) circle parallel to BC and tangent to AB to the
product of the diameters of the two circles parallel to BCand passing by D
and E.
In Figure 15, we have reproduced the page containing the last three propo-
sitions (89 to 91) from Ms. 3464; the reader can notice again the conciseness
of the writing. Let me mention in this regard that I know of no written proofs
of these three propositions, or of any of the last eleven propositions of the
treatise. Furthermore, all the texts that reached us contain some ambigui-
ties in the statements of these propositions. We have discussed the problems
raised by these propositions in our edition [31].
Figure 15. Propositions 89 to 91 of the Spherics, Ms. 3464,

fol. 99, Topkapi Saray, Ahmet III.
Figure 16. The Sector Figure (Proposition 66)
Group 6: The Sector Figure and its applications; Propositions 66-

70 and 72-80. Proposition 66 is the so-called Sector Figure. establishes a
relation between six sides of a figure called “complete quadrilateral”. This
is the figure obtained by producing two pairs of opposite edges of a convex
spherical quadrilateral until they pairwise meet at a point.13 A picture
of a complete quadrilateral is given in Figure 16, where we start with the
quadrilateral ADEG and we produce AG and DE until they meet at the
point B, and AD and GE until they meet at a the point C. Notice that
each convex spherical quadrilateral gives rise to four complete quadrilaterals,
depending on the choice of two adjacent sides. Using the notation of Figure
16, the proposition says that we have:
crd 2AG crd 2AC crd 2DE
= . .
crd 2BG sin crd 2CD crd 2EB
Expressed in the language of sines—which Menelaus did not have, but
which was used by the Arab commentators of the Spherics—the relation is:
sin 2AG sin 2AC sin 2DE
= . .
sin 2BG sin 2CD sin 2EB
In some sense, this spherical proposition is close to the spherical Sine Rule
for triangles; indeed the Sine Rule can be deduced easily from the Sector
Figure, and conversely, there is a very short proof of the Sector Figure using
the spherical Sine Rule. One should note however that the discovery of the
spherical Sine Rule occurred only in the 11th century. In the meanwhile, the
Sector Figure has been the only trigonometrical tool14 used in astronomy.
In particular, the proposition was heavily used by Ptolemy in the Almagest
as the main ingredient for his astronomical calculations. Furthermore, it is
the only result of the Spherics which is quoted under the name “Menelaus’
Theorem”.15
13Convexity is understood here with respect to the intrinsic geometry of the sphere: A
figure is said to be convex if (1) it is contained in an open hemisphere; (2) for any two
points in this figure, the arc of great circle joining them is contained in the figure.
14We are using here the the word “trigonometry” for a formula that gives relations
between the elements of a figure which is not necessarily a triangle.
15There are other names for this proposition. Ptolemy, in the Almagest, used the
expression “secant figure”, which is also the name that was generally adopted by the
) in their commentaries on Menelaus’ and Ptolemy’s works and in the
Arabs ( ¨A¢ ®Ë@ É¾Ë@
treatises they wrote on spherical geometry.
Proposition 66 occupies a particular place in the Spherics. It is situ-

ated among a series of propositions whose proofs are completely intrinsic
to the surface of the sphere, but the proof given to this proposition in al-
Harawı̄’s redaction and in all the other Arabic manuscrits that survive is,
like the one given by Ptolemy in his Almagest, purely Euclidean. (In fact,
the proofs given by the Arabs are either Ptolemy’s or a small variation it.)
At the same time, such a proof was considered as faulty by the various Arab
commentators and editors of the Spherics because it is not in the spirit of
Menelaus’ methods. Whether Menelaus had a proof of this proposition that
does not rely on solid Euclidean geometry remains a mystery.
The other propositions in this group are proved using the Sector Figure.
They involve divisions of spherical triangles by arcs of great circles joining
vertices to the sides that subtends them. These results concern angle and
external angle bisection (Propositions 73 and 79), angle division into three
equal parts (Propositions 74 and 75), division by altitudes (Propositions
78, 80), and various other kinds of divisions. These propositions may all
be proved using the spherical Sine Rule, and the Sector Figure Theorem is
used in the absence of this theorem. We have discussed in detail all these
propositions in our commentary in [31]. In particular, Proposition 80 needs
a careful discussion, and we comment on it here. In al-Harawı̄’s version, the
proposition says the following:
The altitudes that are produced from the angles of a triangle to their op-
posite sides intersect in one point.
The statement as it stands is not correct and needs a hypothesis. Let me
explain this. Consider the example of a spherical triangle whose three angles
are right. In this case, any arc of great circle drawn from a vertex to the
opposite side that subtends it is an altitude. Obviously, three arbitrary such
arcs drawn from the three vertices to the opposite sides do not necessarily
intersect in a common point. Thus, the theorem fails in this case. There are
much more general cases where the theorem fails.
The statement becomes correct if we add to it a further hypothesis. One
possibility is the following:
Suppose that in a given triangle two altitudes intersect in a point within
the triangle. Then there exists a third altitude passing through this point.
With such a hypothesis, the proof in al-Harawı̄’s edition of Menelaus, of
the fact that the three altitudes intersect in one point, becomes correct. This
proof involves several applications of the Sector Figure Theorem.
Group 7: Proposition 71. I have dedicated a special section to Propo-

sition 71 because I consider it as very special, both mathematically and
historically, and I will explain why.
First, the statement is particularly interesting. It expresses a non-trivial
property of spherical triangles, and I have not seen anywhere else, in the old
and modern literature, a statement that bears any resemblance to it. Sec-
ondly, the proof of this proposition has given rise to an abundant literature
by the Arab mathematicians who have revised and commented Menelaus’
text. It seems that it was precisely in trying to prove this proposition that
these mathematicians discovered the Sine Rule and the other formulae of
spherical trigonometry, see the comments in [31] and Section 7 below. More-
over, the proof uses an important property that was obviously known to
Menelaus, namely, the invariance of the spherical cross ratio, although he
used it without proof. The invariance of the cross ratio is one of the main
tools of projective geometry.
I have given a detailed proof of this proposition and reviewed several
points in the history of the important mathematical activity that took place
around it, between the ninth and the thirteenth centuries, in my two papers
with Rashed [29] and [30] .
Let me recall the statement of this proposition:
Consider two triangles ABE and LM O with right angles at B and M ,
and where the angles A and L are acute and equal. Then,
crd (AE + AB) crd (LO + LM )
= .
crd (AE − AB) crd (LO − LM )
I have reproduced the figure that accompanies this proposition in Figure

17 here.
As a matter of fact, inspired by this kind of result in spherical geometry,
several theorems in hyperbolic geometry are proved in the recent paper
[25]. In particular, a more precise analogue of Proposition 71, in the setting
of hyperbolic geometry, is given there (Theorem 2.1), based on hyperbolic
trigonometry. Going backwards, a more precise form of the same proposition
in the setting of spherical geometry can be given using the same outline, and
we have reproduced it in the first appendix to the present paper. Let me
also mention here that Menelaus’ theorem as well as the invariance of the
cross ratio which is used in the proof of Proposition 71 have been adapted
and used in the case of hyperbolic geometry, see [25, 26]. I am saying all this
to give a hint of the freshness of the ideas behind this and other propositions
of the Spherics.
5. Pappus
Pappus of Alexandria is among the rare mathematicians from Greek An-
tiquity who have embraced Menelaus’ intrinsic methods. Book VI of his
Mathematical collection contains a sequence of theorems on spherical geom-

etry, some of which are proven using the intrinsic geometry of surfaces, in
the pure style of Menelaus. As a matter of fact, the first proposition in
this book concerns the triangle inequality, for a spherical triangle. Pappus’
statement is the following:
If, on a spherical surface, three arcs of great circles which are all smaller
than the semi-circle mutually cut each other, then two of these arcs, taken
in an any way, are greater than the remaining arc.
The reader may notice that there is no mention of spherical triangle in
this statement, but right after the proof of this proposition, Pappus notes:
“Menelaus, in his Spherics, calls such a figure a trilateral (τρίπλευρον)”. Pap-
pus uses this word in the subsequent propositions. Another remark that fits
here is that whereas Menelaus’ proof of the triangle inequality is purely
intrinsic, Pappus’ proof uses the ambient Euclidean solid geometry (it is
based on Proposition XI.20 of the Elements, which is a triangle inequality
that concerns the three dihedral angles that form a solid angle.) The reason
for which Pappus’ proof must necessarily use Euclidean geometry is that
this is the first proposition he gives on spherical geometry. As we already
said, regarding Menelaus, spherical geometry, for both authors, based on the
axioms of Euclidean geometry, and the first proposition, in any case, must
use the Euclidean setting.
But the proofs of the next propositions of Pappus are intrinsic to the
sphere. The statements are not always equivalent to those of Menelaus, and
some of Pappus’ propositions complement some propositions of his prede-
cessor.
Let us now mention a few propositions of Pappus.
Proposition 2 says the following [27, p. 371]:
If two arcs of great circles are drawn internally, above one side of a tri-
lateral, then they are smaller than the remaining two sides of the trilateral.
In other words, the proposition says that if we take an interior point in a
triangle and if we join by two arcs of great circles that point to the vertices
of the base, then the sum of these two arcs is smaller than the sum of the
two legs of the triangle. This proposition is equivalent to Proposition 6 of
the Spherics whose statement is [31, p. 518]:
If we produce from two extremities of a side among the sides of a triangle
two arcs of great circles that meet within the triangle, then they are smaller
than the remaining sides.
Then, follow 9 propositions which are proved using Menelaus’ methods.
Some of the statements are close to propositions in Menelaus’ Spherics,
with slightly different hypotheses regarding the bounds on the lengths of
the sides, and some other statements are contained in Theodosius Spherics
(with proofs using Menelaus’ intrinsic methods). As an example of Pappus’
intrinsic methods, let us consider his proposition VI.4, which is equivalent to
Proposition 38 of Menelaus (with a slightly different hypothesis concerning
the lengths of the sides). The latter belongs to Group 4 of the Spherics,
concerning the division of a triangle.
Menelaus’ statement is the following [31, p. 618]:
Figure 18. The figure for Pappus’ Proposition VI.4
If the sum of two unequal sides in a triangle is smaller than a semicircle,

if we separate from the two extremities of the remaining side two equal seg-
ments, and if we produce from the two resulting points two arcs till the angle
that the two sides contain, then they contain with the two sides two unequal
angles, the greater at the smaller and the smaller at the greater, and the two
produced arcs are smaller than the two sides of the triangle.
Pappus’ statement is the following [27, p. 373]:
Let four arcs of great circles AB, AC, AD, AE cut an arc of great cir-
cle BE, let the arc BC be equal to the arc DE, and let each of the arcs
AB, AC, AD, AE be smaller than a quadrant. Then, the sum of the arcs
BA, AE is greater than the sum of the arcs CA, AD.
We reproduce Pappus’ proof in order to show the method (see Figure 18).
Let us cut the arc CD into two equal parts at the point Z. Let us draw
through the points A, Z the arc of great circle AZH, and let us set the arc
ZH to be equal to the arc AZ. Let us draw through the points H, E the
great circle HEK and through the points H, D the great circle HDΘ. Since
the arc HZ is equal to the arc ZA and the arc DZ is equal to the arc ZC,
it follows that the arc DH is also equal to the arc CA. For the same reason,
the arc EH is also equal to the arc BA. But since the two arcs AD, DH
are established internally on one of the sides HA of the trilateral HEA, it
follows that the arcs AD, DH are smaller than the arcs AE, EH. Thus, the
arcs AE, EH are greater than the arcs AD, DH. But the arc EH is equal
to the arc AB and the arc HD is equal to the arc AC. Therefore, the sum
of the arcs BA, AE is greater than the sum of the arcs CA, AD. This is
what we had to prove.
I know of no other mathematician than Pappus before the modern times

(that is, before the nineteenth century) who integrated in his writings Menelaus’
methods of intrinsic geometry of the sphere.
6. Translations, rectifications and difficulties

Everything we know about the history of the translations and the various
versions of Menelaus’ Spherics is reported on in the first part of the edition
[31], p. 3-123 and in the two articles [29, 30]. I will summarize here some
important points on that history.
Several Arabic versions of the Spherics were written starting in the 9th
century. Al-Harawı̄’s version, which relies on a version by al-Māhānı̄ and
which is a rectification16 rather than a translation, is published for the first
time, together with an English translation, in the volume [31]. Besides al-
Harawı̄’s version, this volume contains the first critical edition of a fragment
containing the 36 first propositions of a translation done in the same century,
which we think is the oldest translation that survives. We know that another
translation was also done in the same century, by the Arab mathematician
and Hellenist Abū ‘Uthmān al-Dimashqı̄. A Latin translation by Gerard of
Cremona (1114-1175) and a Hebrew one by Jacob ben Makhir Ibn Tibbon
(ca. 1273) were done from Arabic texts. According to the latter, there
existed another Arabic translation by Ishāq ibn H . unayn, or by his father
H. unayn ibn Ishāq. Besides these translations, there are also rectifications
by Abū Nas.r Mans.ūr Ibn ‘Alı̄ Ibn ‘Irāq (born around the year 950), Nas.ı̄r al-
Dı̄n al-T.ūsı̄ (13th c.), Ibn Abı̄ Jarrāda (13th c.), Muh.yı̄ al-Dı̄n al-Maghribı̄
(13th c.), al-Yazdı̄ (15th c.), and certainly others. A German translation of
the rectification by ‘Irāq was published in 1936 by M. Krause, see [20]. These
versions differ on several points, with additions of lemmas and intermediate
propositions, etc., but the overall mathematical content is the same. There
are also writings and developments based on Menelaus’ Spherics by Ibn Hūd,
Thābet ibn Qurra, al-Bı̄rūnı̄, al Qiftı̄, Ibn al-Haytham and others. We have
discussed all this in the volume [31] where we have also made comparisons
between significant passages of the existing Arabic versions that flourished
from the 9th century on and continued till the thirteenth, and the Latin and
Hebrew versions.
We have quoted, in the introduction of the present paper, an excerpt
of the Preamble of al-Harawı̄’s rectification, in which the latter expresses
his wonder and admiration for Menelaus’ geometric reasoning, which he
considers as completely new. In the same Preamble, al-Harawı̄ mentions
the difficulties inherent to the various translations, due to the fact that the
geometers of his time were not capable of understanding several propositions
and proofs of this treatise. He writes [31, p. 24]:
16the Arabs used the words Is̄lāh (

. hC @ ), which means “reform”, or “revision” (an
example is the treatise by Abū Nas.r Mans.ūr ibn ‘Alı̄ ibn ‘Irāq, Is̄lāh. Kitāb Mānālāwus fı̄

al-ashkāl al-kuriyya, Leiden, Or. 930), and Tah.rı̄r ( QK
Qm' ), which means “redaction” (an
example is Nas̄ı̄r al-Dı̄n al-T
. ūsı̄’s Tah.rı̄r Kitāb Mānālāwus fı̄ al-ashkāl al-kuriyya [Hyder-
abad, 1359 H]).
When, from that book, the geometers received what they were
not used to, and to nothing similar from all other kinds of ge-
ometries they knew, they found it difficult. The bad translation
and the fact that it was far from being understandable added to
their apprehension, so they discarded it and they did not work
on it. Nevertheless, no one among them denied the eminence of
that book, but they acknowledged that what he showed in this
way is easier and more valuable that anything which anybody else
invested in.
It was said that a group of geometers had the intention to rectify
it, and since they were unable to do it, they asked the help of al-
Māhānı̄, who rectified the first book and some propositions of the
second, and he stopped at a proposition of which they said that
its aim was difficult, and whose explanation was also hard.
[. . . ] I meditated on what al-Māhānı̄ rectified and I saw during
this short period of time that there were defects, and I carefully
reviewed what needed to be corrected in the expression, in the
meaning, and in the proof. And I persisted until the end of the
book, until the proposition which the geometers imagined that
al-Māhānı̄ left aside and was not able to rectify. I also found a
rectification that was far from being correct due to some modern
geometer who claimed he rectified part of it, and left aside the rest.
In the thing he said that he rectified there are some corrupt things,
from which it was clear to anybody who examined them that he
did not understand the purpose of this man [i.e. Menelaus].
According to the testimony of Nas.ı̄r al-Dı̄n al-T.ūsı̄, the situation of the
existing versions of the Spherics in the thirteenth century was not better,
except for the one by Ibn ‘Irāq. He writes:17
When I reached the book of Menelaus on the spherical figures, I
found it in many different copies, with undetermined problems,
together with disappointing rectifications, like the rectifications of
al-Māhānı̄ , of Abı̄ al-Fad.l Ah.mad ibn abı̄ Sa‘ad al-Harawı̄ and oth-
ers, some of which are incomplete and others wrong. I was puzzled
for what concerns the proof of some problems, until I obtained the
rectification of Abū Nas.r Ibn ‘Irāq.
It is likely that Ibn ‘Irāq’s is a rectification of the translation due to the
Hellenist and mathematician Abū ‘Uthmān al-Dimashqı̄. We refer again the
reader to our discussion in [31].
In addition to his redaction of the Spherics, Ibn ‘Irāq wrote a short treatise
which concerns Proposition 71 which we discussed above (in our classification
on §3, we included it in Group 8), titled The Rectification of a Proposition
of the Book of Menelaus on the Spherics. A translation of this treatise is
provided in the article [29].
Ibn Abı̄ Jarrāda, a contemporaty of Tūsı̄, considered that the most pop-
ular rectification that existed at his time was still that of al-Māhānı̄/al-
Harawı̄, which, despite the fact that, according to him, it has many short-
comings. Let us quote him (see [31, p. 18]):
I found that book (Menelaus’ Spherics) so altered as none of
the other geometry books, for reasons that were specific to it,
17Reference in [31, p. 16]
among which are the difficulties in the proofs, for someone who
was supposed to clarify them, because of the necessity of these
conditions. And among these difficulties is the fact that Menelaus
over-summarized it, as he mentioned in the Preamble. And also,
among these difficulties, the multiplicity of the problems hidden
in the course of the proofs, and of the abandoned ones to which
he refers. Among these, that he meant by the ratio of the arcs
the ratio of the double of their chords in the entire third book and
in part of the fourth, and he meant by the ratio of the arcs their
own ratio in other parts of the fourth book, thinking, because of
his extreme power in that field, that this would not be hidden to
anyone who will undertake the clarification of his book. That was
his basis, and it became difficult to anyone who came after him to
recognize this. Part of <the difficulties> is also the fact that he
refers in the proofs of the propositions of the fourth book to what
he has proved in one of his books, titled The Measurable Figures,
a book which is unknown, or else it is foreign and did not reach
us.
A group <of people> proposed to rectify this book, all of them
had failed the aim, they stopped before reaching their goal and they
did not comply with Menelaus’ method. Among them are those
who were not able to penetrate it. There are also those who rec-
tified one part of the book, in the given order, and were incapable
of going through it, like al-Māhānı̄, who reached a proposition
which he was incapable of solving, and thus he did not complete
it. Among them are also those who pretended they have rectified
all the book, deluding and disappointing anyone who examines it.
They relied on the fact that when one encounters a difficulty in
understanding it, he attributes this lack either to himself or to the
corruption of the transcription.
Even though [al-Harawı̄’s] rectification is the most famous one

among those which the book was subject to, the best that he brings
is where he limits himself to repeat Menelaus’ expression. But as
soon as he proposes himself to prove a neglected <problem> or to
clarify a difficulty, or something similar, then he leaves the condi-
tions of Menelaus, like in the first lemma by which he proved the
proposition where al-Māhānı̄ stopped, proving it with common sec-
tions whereas what he brought does not suffice for the completion
of that proposition. Or as he did in the thirty-seventh proposi-
tion of the first book: he resorted to a reductio ad absurdum for
what is hidden in the proof of Menelaus. Regarding the fourth
book, which is the third in his rectification, he did not understand
anything of it except the ninth proposition and he attributed ev-
erything which is in that book concerning the ratios of arcs to the
ratios of chords of the double. Everything that I mentioned should
be clear to anyone who examines the rectification of al-Harawı̄.
This translation and rectification work on Menelaus’ Spherics generated

an important mathematical activity that changed the course of Arabic math-
ematics. This is the subject of the next section.
7. On the impact on Arabic mathematics

Among Menelaus’ propositions that had a major impact on Arabic math-
ematics is the Sector Figure (Proposition 66). Several treatises were written
on the applications of this proposition. The oldest known such complete
treatise is the one of Thābit ibn Qurra (d. 901), titled A Treatise on the
Sector Figure, is reviewed and commented on in detail by H. Bellosta in her
article [3]. We quote from the Introduction to this treatise [3, p. 365]:
I know that people never discussed any geometrical proposition
used in astronomy as much as they did for this proposition. They
did not overrate its reputation and the reason of their interest for
it; this is what we know from the number of times it was used and
the extreme need that we have for it in the spherical science. This
is certainly the main direction on which many works in astronomy
are based. Even though others than Ptolemy obtained this propo-
sition and talked about it, nobody, to our knowledge, improved its
statement or <gave> any arrangement or revision that matches
its proof.
Other developments concerning the Sector Figure were reported on by
Abū al-‘Abbās al-Nayrı̄zı̄ and Abū Ja‘far Muh.ammad ibn al-H . usayn al-
Khāzin, in their commentaries of the Almagest. There is a large literature
on the impact of this proposition on Arabic mathematics. The reader may
refer to the work of M.-Th. Debarnot [9, 10], and the latter’s edition of the
10th century work of al-Bı̄rūnı̄ [11]. We have commented on this proposi-
tion and its applications in the commentary contained in our edition of the
Spherics; see [31, p. 292–337].
In the 13th century, Nas.ı̄r al-Dı̄n al-T.ūsı̄, wrote a treatise on the Sector
Figure [35] in which he showed how a complete trigonometric system of
spherical geometry can be derived from this proposition. Al-T.ūsı̄’s treatise
was translated into French and edited by Alexandre Carathéodory18 in 1891
[35]. See the detailed review in [23]. It is only in the eighteenth century that
Leonhard Euler published a treatise containing such a complete system,
obviously, without being aware of al-T.ūsı̄’s work, see [15]. Nas.ı̄r al-Dı̄n
uses in his work the notion of polar triangle. This notion and the spherical
Sine Rule were already introduced in the tenth century by Ibn ‘Irāq and
Abū al-Wafā’ al-Būzjānı̄ (940-998). We have recalled, in [31, p. 795-801],
the story of these discoveries and we have pointed out the confusion that
reigned in the 18th and 19th century literature concerning the history of
spherical trigonometry, in particular concerning the discovery of the polar
triangle.
Nas.ı̄r al-Dı̄n’s treatise is a remarkable historical document that gives a
very good idea of the state of spherical geometry, as developed by the Arabs
between the ninth and the thirteenth centuries. At the end of Book IV,
Nas.ı̄r al-Dı̄n says that the Ancients have used the quadrilateral figure with
confidence, without treating the large number of cases, and he mentions as
examples Menelaus’ Spherics and Ptolemy’s Almagest. But the moderns, he
18Alexandre Carathéodory was a great-uncle of the famous mathematician Constantin

Carathéodory.
says, in order to avoid engaging themselves in the examination of the mul-

titude of ratios and the lengthy considerations that the use of compounded
ratios requires, imagined and studied other figures, which replace the quadri-
lateral, which are as much useful and which do not lead to a multitude of
cases. The presentation of these methods is the subject of Book V of his
treatise.
We already mentioned another theorem of the Spherics that had a large
impact on Arabic mathematics, especially on the development of spherical
geometry with works by al-Būzjānı̄, al-Khujandı̄, Ibn Labbān, Ibn ‘Irāq and
al-Bı̄rūnı̄. This is Proposition 71. As we already recalled, Menelaus did
not give a proof of this proposition, but only an outline of a proof. Writ-
ing a complete proof became a big challenge that was taken up by several
mathematicians, in Baghdad and elsewhere, between the ninth and the thir-
teenth centuries. We already mentioned that the proof of this proposition
uses an invariance property of the spherical cross ratio, and to prove this
invariance, the mathematicians of the tenth century used the spherical Sine
Rule which they had established. It is conceivable that the development of
this rule was motivated by the research in spherical geometry due to the
effort to prove this proposition of Menelaus. In fact, the search for a proof
of this proposition acted as a motivation for reshaping the whole field of
spherical geometry. Likewise, it is possible that the discovery by Ibn ‘Irāq of
the notion of polar triangle was also motivated by his work on the proof of
Menelaus’ Proposition 71. In his Rectification of a Proposition of the Book of
Menelaus on the Spherics, which is edited and translated in the paper [29],
Ibn ‘Irāq presents a set of propositions on spherical geometry that are useful
in the proof of this proposition, but that are also of independent interest.
At the same time, he gives a proof of the proposition under assumptions
that are weaker than the ones that Menelaus made. While he worked on
the proofs, Ibn ‘Irāq declared that not only Proposition 71, but the whole
treatise of Menelaus’ Spherics needs to be corrected. He reported on the
attempts by al-Māhānı̄ and al-Harawı̄ on correcting Menelaus’ treatise and
he discussed several points in their proofs. We have discussed these points
in our article [29].
Understanding the proof of Proposition 71 became an important matter,
not only for mathematicians, but also for astronomers and other scientists.
The last section of our paper concerns astronomy.
8. On the impact on astronomy

In his introduction to his rectification, al-Harawı̄ writes [31, p. 501]:
“Ptolemy relies on this book, especially in the second book of the Almagest
concerning angles and triangles produced by intersections of great circles.”
In the same introduction [31, p. 504], he writes: “This book has an immense
value in astronomy and the properties of the figures that occur when the
horizons intersect the ecliptic, the orbs of the meridian, the equator and
the circles called the ephemerides that the sun draws at the beginning of the
zodiac.” Concerning Menelaus’ Proposition 66 (the Sector Figure), al-Harawı̄
writes [31, p. 684]:
And he worked out in that book the figure which Ptolemy calls the
“sector”, and he constructed upon it several figures. But Ptolemy
uses, from that book, numerous propositions in the second book
of his treatise the Almagest, without attributing them to anybody
and without proving anything. Indeed, all what he uses, for what
concerns angles, arising from the intersection of the ecliptic and
the horizons, and other things, can be proved by this book. The
lemmas we need are among those which Ptolemy introduced, on
the intersection of two lines and on the composition of the ratios
which are made up thereof. And we find that Menelaus moved
in that book to the Sector Figure in a way which does not fit his
method, since he did not set for this passage any lemma or any
discourse and he also did not make it the starting of a book. So
either the lemmas of that figure were known by them, widespread,
or they disappeared from that book.
The astronomer Ibn al-Haytham, who was familiar with the Sine Rule
and who used it, also used the Sector Figure Theorem in his astronomical
works.19
With all this, we must say that the astronomical side of Menelaus’ Spher-
ics has been so much exaggerated to the point that some authors have
considered that the work (or at least part of it) was written for its use in
astronomy. The confusion started with authors like Bjørnbo and Heath who
(mistakenly) declared that a substantial part of the Spherics is astronomical,
or that it has been written for the purpose of astronomy.20 These statements
have been repeated tirelessly and superficially by several modern authors.
The first historians to put forward these ideas were probably convinced that
among the Greeks, the only sphere worth studying was the celestial sphere,
an idea that we, mathematicians, find hard to believe, knowing that the
sphere and its intersections by planes is as natural a field of mathematics to
study, as are the cones and their intersections by planes, which are the sub-
ject of study of Apollonius’ Conics, and, in fact, as the ordinary plane with
its straight lines . It is probable that Heath was influenced in his opinion by
A. A. Bjørnbo [4]. In [31, p. 3] we have given a list of authors from before
the 1950s who repeated the statement that Menelaus’ Spherics (or at least a
substantial part of it) is a work on astronomy. The names of many modern
authors may be added to this list, and we prefer not to dwell on this.
Another statement of Heath which led to a misinterpretation is that he
declared that Book III of Menelaus’ Spherics belongs to “trigonometry”,21 a
statement which is also absurd. Heath writes in [19, p. 265]: “It will have
19See e.g. Proposition 5 of the treatise of Ibn al-Haytham titled The Configuration of
the Motion of Each of the Seven Wondering Stars, pp. 294 sqq., in: Roshdi Rashed, Les
mathématiques infinitésimales du IXe au XIe siècle, vol. V: Ibn al-Haytham. Astronomie,
géométrie sphérique et trigonométrie, London, 2006.
20Heath’s statement, concerning Menelaus’ Spherics, in [19, p. 265], is bewildering:
“Book II has practically no interest for us. The object of it is to establish certain propo-
sitions, of astronomical interest only, which are nothing more than generalizations or
extensions of propositions in Theodosius’ Sphaerica, Book III.”
21In Heath’s treatise [19], the study of Menelaus’ book is part of the chapter called
Trigonometry.
been noticed that, while Book I of Menelaus gives the geometry of the spher-
ical triangle, neither Book I nor Book II contains any trigonometry. This is
reserved for Book III.” In fact, spherical trigonometry has been developed
only several centuries later. The only reasonable meaning that can be given
to Heath’s statement is that Book III (like Books I and II) constitue a study
of spherical triangles, but if such a the study is termed as trigonometry, then
several books of Euclid’s Elements should also be termed as trigonometry.
In fact, Heath says (p. 260) that “Greek trigonometry in its highest devel-
opment [is] in the Spherica of Menelaus,” which is correct, but it is only
a relative statement. It is correct if we mean by this “trigonometry” in its
etymological sense, the study of triangles, but then we have to include also
Books I and II of the work.
On the other hand, taking about Greek astronomy, it is fair to recall that
Theodosius wrote (at least) two astronomical treatises in which the celestial
phenomena are studied from a geometric point of view, namely, On habi-
tations and On the days and nights. The treatise On habitations contains
12 propositions which are all concerned with how the stars in the celestial
sphere are seen from different places of the Earth. A brief paraphrase of the
propositions is contained in Delambre’s Histoire de l’astronomie ancienne,
[12], vol. 1, p. 235-7. I have reproduced in Figure 19 a page containing
Proposition 9 of the Habitations, which concerns a comparison between the
location of the stars that are visible and the duration of the visibility of
these stars, at two places that are not situated on the same meridian. For a
recent edition of the Arabic translation of the On habitations, due to Qust.a
ibn Lūqā appeared in 2011, together with the Latin translation (done from
the Arabic) by Gerard of Cremona and an English translation, see [34].
This being said, I would like to to stress, first, on the fact, that Heath
remains one of the most valuable specialists of Greek mathematics in modern
history, and secondly, that, by an ironic twist of history, it is true that
Ptolemy, in his great astronomical treatise, relied heavily on a proposition
of Menelaus’ Spherics, which is not an astronomical work, rather than on the
astronomically oriented Habitations or Days and nights of Theodosius, or on
the Moving sphere or the Risings and settings of Autolycus, the reason being
of course that this proposition of Menelaus could be used in an effective way
in astronomical calculations.
9. As a word of conclusion
In the preceding pages, I have talked about the history of Menelaus’ Spher-
ics, discussing its relation with ancient and modern research on geometry.
In the next two appendices, I have included modern proofs of two propo-
sitions of Menelaus, using spherical trigonometric formulae. Such modern
proofs are straightforward and they sometimes give more precise statements
than the synthetic arguments. However, there is an enormous chasm be-
tween these automatic proofs and the depth of Menelaus’ geometric proofs,
in which the etymology of the word Theorem 22 as a contemplation finds it
full force. It is in such instances which after all are rare in the history of
mathematics that one can recognize the seal of a real mathematician.
22From the Greek, θεωρέω, “I watch”, the same origin than the word theatre.
Figure 19. From Theodosius’ Habitations; at the bottom

of the page, the statement of Proposition 9 and the corre-
sponding figure. Version of Qust.a ibn Lūqā, Ms. 3464, fol.
121, Topkapi Saray, Ahmet III.
Appendix 1: A proof of Proposition 27 by spherical

trigonometry
We start with a formula of Euler from his memoir [14] which says that
the area of a spherical triangle ABC in terms of the lengths of sides a, b, c
is equal to:
1 + cos b + cos a + cos c

.
4 cos 2b cos a2 cos 2c
Consider now a triangle ABC, let D and E be the midpoints of AB and
BC, and DE the arc of great circle joining them.
The cosine formulae in the triangles ABC and ADE give:
cos a cos c − cos b
cos B =
sin a sin c
and
cos a2 cos 2c − cos DE
cos B = .
sin a2 sin 2c
Equating the two values of cos B, we obtain
(cos b − cos a cos c)(sin a2 sin 2c ) + cos a2 cos 2c sin a sin c
cos DE = .
sin a sin c
Replacing in the right hand side cos a and cos c by their values in terms of
sin a2 , sin 2c , cos a2 and cos 2c , and after some cancellations, we get:
cos b + 2 cos2 a2 + 2 cos2 2c − 1
cos DE =
4 cos a2 cos 2c
cos b + (1 + cos a) + (1 + cos c) − 1
=
4 cos a2 cos 2c
1 + cos b + cos a + cos c
=
4 cos a2 cos 2c
1 + cos b + cos a + cos c b
= b a c
. cos
4 cos 2 cos 2 cos 2 2
Using Euler’s formula we get
1 b
cos DE = cos Area(ABC) cos ,
2 2
which implies DE < AC/2.
Appendix II: A proof of Proposition 71 by spherical

trigonometry
The following formula gives a more precise version of Menelaus’ proposi-
tion 71:
Theorem.— Given a spherical triangle ABC with right angle at A and
acute angle at C, and such that AC is smaller than a quarter of a circle,
then we have:
sin(BC + AC) 1 + cos C
= .
sin(BC − AC) 1 − cos C
In particular, the ratio of sines in the conclusion depends only on the

angle C, which implies the result of Proposition 71.
The proof is given in the setting of hyperbolic geometry in the paper [25].
That proof is based on some non-Euclidean trigonometric transformations.
It is easily adapted the spherical case, replacing the hyperbolic functions
sinh and cosh by the functions sin and cos. We use the following lemma,
which follows from the spherical cosine and Sine Rules.
Lemma.— We have
tan BC = cos C · tan AC.
Then we expand sin(BC + AC) and sin(BC − AC):

sin(BC + AC) sin BC cos AC + sin AC cos BC
=
sin(BC − AC) sin BC cos AC − sin AC cos BC
sin AC sin BC
cos AC + cos BC
= sin BC sin AC
.
cos BC − cos AC
Using the lemma, the last term becomes
1 + cos C
,
1 − cos C
which is what we wanted to prove.
References
[1] Autolycos de Pitane, La sphère en mouvement, Levers et couchers héliaques, Testi-
monia, éd. Germaine Aujac, avec la collab. de J.-P. Brunet et R. Nadal, Paris, Les
Belles Lettres, 1979.
[2] P. Barbarin, La géométrie non euclidienne, Paris, C. Naud, 1902.
[3] H. Bellosta, Le traité de Thābit ibn Qurra sur la figure secteur, in R. Rashed
(ed.), Thābit ibn Qurra: Science and Philosophy in Ninth-Century Baghdad, Sci-
entia Graeco-Arabica 4, Berlin / New York, 2009, pp. 335-390.
[4] A. A. Bjørnbo, Studien Ã 14 ber Menelaos’ Sphärik: Beiträge zur Geschichte der
Sphärik und Trigonometrie der Griechen, Leipzig, 1902
[5] Busemann, H. (1948). Spaces with non–positive curvature. Acta Math. 80: 259–311.
[6] Busemann, H. (1955). The Geometry of Geodesics. Pure Appl. Math. 6. New York:
Academic Press Inc.
[7] Busemann, H. (1961). The Synthetic Approach to Finsler Spaces in the Large. In:
Geometria del calcolo delle variazioni. C.I.M.E. Summer Schools, vol. 23. Berlin,
Heidelberg: Springer.
[8] Clavius’s commentary on the spheriks of Theodosius Tripolitæ: or, Spherical ele-
ments, necessary in all parts of mathematics, wherein the nature of the sphere is
considered, English transl. E. Stone, London, Senex, Taylor and Sisson, 1721.
[9] M-Th. Debarnot, Trigonometry, in R. Rashed (ed.), Encyclopedia of the History of
Arabic Science, 3 vols., London and New York, 1996, vol. II, p. 160-203.
[10] M-Th. Debarnot, Trigonometria, in Storia della scienza, vol. III: La civiltà islamica,
Roma, Enciclopedia Italiana, 2002, p. 432-447.
[11] M.-Th. Debarnot (ed.), Al-Bı̄rūnı̄, Kitāb Maqālı̄d ‘ilm al-hay’a: La trigonométrie
sphérique chez les Arabes de l’Est à la fin du Xe siècle, ed. Institut FranÃ§ais du
Proche-Orient, Damascus, Vol. 114, 1985, 314 p.
[12] J.-B. Delambre, Histoire de l’astronomie ancienne, Courcier, Paris, 1817.
[13] L. Euler, Principes de la trigonométrie sphérique tirés de la méthode des plus grands
et plus petits, Mémoires de l’académie des sciences de Berlin, Volume 9, p. 223-257,
1753, Opera Omnia, Series 1, Volume 27, p. 277-308.
[14] Euler, L. (1797) Variae speculationes super area triangulorum sphaericorum. Nova
Acta Academiae Scientarum Imperialis Petropolitinae. 10 (1797), 47–62. In: Opera
Omnia, Ser. 1, Vol. 29, 253–266. English translation by R. Caddeo, in Spherical
geometry in the eighteenth century: Euler, Lagrange and Lambert, Springer, 2013, to
appear.
[15] L. Euler, Principes de la trigonométrie sphérique tirés de la méthode des plus grands
et des plus petits, Mémoires de l’Academie des Sciences de Berlin 9 (1753), 1755, p.
233-257; Opera Omnia, Series 1, vol. XXVII, p. 277-308. English translation by A.
Papadopoulos and A. Zhukova in Spherical geometry in the eighteenth century: Euler,
Lagrange and Lambert, Springer, 2013, to appear.
[16] L. Gérard, Sur la géométrie non euclidienne, Thèse de doctorat présentée à la Faculté
des Sciences de Paris, Paris, Gauthier-Villars, 1892.
[17] N. Halma (ed.), Composition mathématique de Claude Ptolémée, traduite par Nicolas
Halma et suivie des notes de Jean Baptiste Delambre, 2 vols., Paris, Henri Grand,
1813. Reprint, 1988, Librairie Blanchard, Paris.
[18] G. B. Halsted,The elements of geometry, New York: J. Wiley and sons, 1885.
[19] Th. L. Heath, A History of Greek Mathematics, vol. II: From Aristarchus to Dio-
phantus, Oxford, 1921; repr. Dover, 1981, p. 265.
[20] M. Krause, Die Sphärik von Menelaos aus Alexandrien in der Verbesserung von Abū
Nas.r Mans.ūr b. ‘Alı̄ b. ‘Irāq, Abhandlungen der Gesellschaft der Wissenschaften zu
GÃ¶ttingen, phil.-hist. Klasse, 3, 17, Berlin, 1936.
[21] J. Lexell, Solutio problematis geometrici ex doctrina spaericorum, Acta academiae
scientarum imperialis Petropolitinae, 5 (1) 1784, p. 112-126. Solution of a geomet-
ric problem according to the spherical theory, English translation by R. Caddeo in:
Spherical geometry in the Eighteenth century (R. Caddeo, A. Papadopoulos and A.
Zhukova ed.), Springer Nature, 2022.
[22] M. Mersenne, Universae geometriae mixtaeque mathematicae synopsis, et bini re-
fractionum demonstratarum tractatus. Studio & opera F. M. Mersenni, 1626. New
augmented edition, 1624.
[23] A. Papadopoulos, Nas.ı̄r al-Dı̄n al-T . ūsı̄’s treatise on the quadrilateral: The art of
being exhaustive, Gan.ita Bhārātı̄, (Indian Mathematics) the Bulletin of the Indian
Society for History of Mathematics, Vol. 40, No. 1 (2018), p. 153-180.
[24] A. Papadopoulos, Three Theorems of Menelaus. Amer. Math. Monthly, 126 (2019),
no. 7, 610–619.
[25] A. Papadopoulos and W. Su, On hyperbolic analogues of some classical theorems
in spherical geometry. In: Hyperbolic geometry and geometric group theory (ed. K.
Fujiwara, K. Ohshika and S. Kojima), Advanced Studies of Pure Mathematics, No.
73, Mathematical Society of Japan, Tokyo, 2017, p. 225-253.
[26] A. Papadopoulos, and S. Yamada, The Funk and Hilbert geometries for spaces of
constant curvature, Monatshefte für Mathematik 172, 1 (2013) p. 97-120.
[27] La Collection mathématique de Pappus d’Alexandrie, translation with notes by P.
Ver Eecke, Paris, Librairie Albert Blanchard, 1982.
[28] R. Rashed, Ibn al-Haytham, Les Mathématiques infinitésimales du IX e au XI e siècle,
vol. V : Ibn al-Haytham : Astronomie, géométrie sphérique et trigonométrie, London:
Al-Furqān Islamic Heritage Foundation, 2006. English translation by J. V. Field, New
Spherical Geometry and Astronomy, A History of Arabic sciences and mathematics,
vol. 4, London and New York, 2014.
[29] R. Rashed and A. Papadopoulos, On Menelaus’ Spherics III.5 in Arabic mathematics,
I: Ibn ‘Irāq, Arabic Science and Philosophy, vol. 24 (2014), p. 1-68.
[30] R. Rashed and A. Papadopoulos, On Menelaus’ Spherics III.5 in Arabic mathematics,
II: Nas.ı̄r al-Dı̄n al-T . ūsı̄ and Ibn Abı̄ Jarrāda, Arabic Science and Philosophy, vol. 25
(2015), p. 1-32.
[31] R. Rashed and A. Papadopoulos, Menelaus’ Spherics: Early Translation and al-
Māhānı̄/al-Harawı̄’s Version, De Gruyter, Series: Scientia Graeco-Arabica, 21, 2017,
890 pages.
[32] H. Senapati, Monotonicity in spherical and hyperbolic triangles, In: Eighteen Essays
in Non-Euclidean Geometry, European Mathematical Society, 2019.
[33] Theodosius, Les Sphériques de Théodose de Tripoli, Œuvres traduites pour la pre-
mière fois du grec en franÃ§ais avec une introduction et des notes par P. Ver Eecke,
Bruges, Desclée de Brouwer, 1927; repr. Paris, Librairie Albert Blanchard, 1959.
[34] Theodosius, De habitationibus: Arabic and Medieval Latin Translations, ed. P. Ku-
nitzsch and R. Lorch, Bayerische Akademie der Wissenschaften, phil.-hist. Klasse,
Sitzungsberichte, 2011, Heft i; C.H. Beck, Munich, 2011.
[35] Nas.ı̄r al-Dı̄n al- T
. ūsı̄, Traité du quadrilatère, attribué à Nassiruddin-el-Toussy, d’après
un manuscrit tiré de la bibliothèque de S. E. Edhem Pacha, ancien grand vizir, traduit
par Alexandre Pacha Carathéodory, ancien ministre des affaires étrangères, Con-
stantinople, Typographie et Lithographie Osmanié, 1891.
Athanase Papadopoulos, Université de Strasbourg and CNRS, 7 rue René

Descartes, 67084 Strasbourg Cedex, France
E-mail address: athanase.papadopoulos@math.unistra.fr

Athanase Papadopoulos

Uploaded by

Copyright:

Available Formats

Athanase Papadopoulos

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

Athanase Papadopoulos

Uploaded by

Copyright:

Available Formats

MENELAUS’ SPHERICS IN GREEK AND ARABIC

MATHEMATICS AND BEYOND

Abstract. We present the history of Menelaus’ Spherics (2nd c. AD), ex-

Keywords. Menelaus of Alexandria - Spherics - Spherical geometry -

AMS codes. 01A20 - 01A30 - 5303 - 53C70

Date: February 21, 2023.

Acknowledgements This work obviously owes a lot to Roshdi Rashed

2. Two predecessors: Theodosius and Autolycus

implicitly to Theodosius’ work, and uses some of his propositions. We have

of the angle made by two planes is contained in Euclid’s Book XI (Definition

Figure 1. From Theodosius’ Spherics. On the top, the fig-

Figure 2. A plate of figures for Book II of Theodosius’

Figure 3. From Theodosius’ Spherics, Proposition II.15.

the culmination of both Spherics, in terms of degree of difficulty. Let me

Figure 4. From Theodosius’ Spherics, Proposition III.10.

Theodosius remains, to the best of our knowledge, the first mathematician

albeit the majesty of what he was undertaking, he bypassed nu-

6“By the direct method” means without using a proof by contradiction.

Some propositions of Autolycus’ Moving sphere are used in Theodosius’

3. Menelaus’ Spherics: Preliminary remarks

Although al-Harawı̄’s rectification of the Spherics is divided into two

Figure 5. Figure for Proposition 41

Group 1: Construction of angles; Propositions 1, 41-44. I start with

Figure 6. Proposition 41 fails with the assumption that AB

Figure 7. Proposition 41 fails when AB and AC are greater

The proof is contained in [31, p. 628-630]. I have reproduced here

Figure 9. Propositions 42, 43 and 44. Ms. 3464, fol. 88,

Figure 10. Proposition 11

Figure 11. Proposition 12

Figure 12. Proposition 27

Group 4: Division of triangles; Propositions 27-40. This group con-

Since DG = DE, AD = DB and EDB \ = ADG, \ the two triangles EDB

Group 5: Monotonicity; Propositions 45-62, 64, 65, 81-91. This

Figure 13. Proposition 82

There is a progression in the results of Group 5 in terms of the sophistica-

Figure 14. Proposition 89

Figure 15. Propositions 89 to 91 of the Spherics, Ms. 3464,

Figure 16. The Sector Figure (Proposition 66)

Group 6: The Sector Figure and its applications; Propositions 66-

Proposition 66 occupies a particular place in the Spherics. It is situ-

Group 7: Proposition 71. I have dedicated a special section to Propo-

Figure 17. Proposition 71

I have reproduced the figure that accompanies this proposition in Figure

Mathematical collection contains a sequence of theorems on spherical geom-

Figure 18. The figure for Pappus’ Proposition VI.4

If the sum of two unequal sides in a triangle is smaller than a semicircle,

I know of no other mathematician than Pappus before the modern times

6. Translations, rectifications and difficulties

16the Arabs used the words Is̄lāh (

Even though [al-Harawı̄’s] rectification is the most famous one

This translation and rectification work on Menelaus’ Spherics generated

7. On the impact on Arabic mathematics

18Alexandre Carathéodory was a great-uncle of the famous mathematician Constantin

says, in order to avoid engaging themselves in the examination of the mul-

8. On the impact on astronomy

Figure 19. From Theodosius’ Habitations; at the bottom

Appendix 1: A proof of Proposition 27 by spherical

1 + cos b + cos a + cos c

Appendix II: A proof of Proposition 71 by spherical