7-2-48. Felix Klein to H. Poincaré, English translation

Leipzig 19 June 1881

Dear Sir,

Upon receiving your welcomed letter yesterday, I immediately sent you those offprints I still have of works relating to our topic. Allow me to add today a few lines of explanation pertaining to them. Of course, a single letter will not suffice; rather must we correspond more extensively in order to establish mutually close contact. Today I would like to put forward the following points:

1. Of the papers sent to you the three most important from the 14th Volume of the Math. Annalen are missing, as are also my investigations of the icosahedron in Volumes 9 and 12, and my second memoir on linear differential equations (which also appears unknown to Mr. Picard) in Volume 12. I entreat you to procure them somehow. I have been sending various offprints to Paris – to Hermite, for instance.

2. The work of my students Dyck and Gierster complements my own. I am asking both of them to send you their offprints. Mr. Hurwitz’s doctoral dissertation, relating to these same theories, is in press and will be delivered to you in a few weeks.

3. Since last Autumn one of your countrymen, whose name you must surely know since he studied under Picard and Appell, namely Mr. Brunel (address Liebigstraße 38/II), has been here. Perhaps you would be interested in starting a correspondence also with him; he could recount – better than myself – the organization of our seminar and the role played therein by single-valued functions invariant under linear transformations.

4. I let Mr. Gierster write up a set of notes from the course I gave in the summer of 1879. At the moment it is on loan, but I should get it back in a few days and will go through it with Mr. Brunel before getting back to you with an account of it.

5. I reject the name “Fuchsian functions” although I understand that you were led to these ideas via Fuchs’s work. In essence, all these investigations are based on those of Riemann. My own evolution in this regard was strongly influenced by Schwarz’s deliberations, closely linked to Riemann’s and of great significance, appearing in Volume 75 of Borchardt’s Journal (and which I can strongly recommend if you are unaware of them). Mr. Dedekind’s memoir on elliptic modular functions appeared only in Volume 83 of Borchardt’s Journal, when the geometric representation of modular functions was already clear to me (by the autumn of 1877). By their ungeometric form, Fuchs’s memoirs stand in deliberate opposition to these works. I don’t deny the great service that Mr. Fuchs has rendered other parts of the theory of differential equations, but his work here leaves so much to be desired that on the only occasion when, in a letter to Hermite, he expatiated on elliptic modular functions, he made a fundamental mistake, which Dedekind only gently criticizes in the above-mentioned memoir.

6. One may, in particular, define a function invariant under linear transformations by the property that it maps the half-plane onto a given polygon with sides circular arcs. This in fact represents only a special case of the general situation (I don’t know yet if you have been limiting yourself to just this particular case). The corresponding group of linear transformations is then characterized by the fact that it is contained in a group of operations twice as large, which, in addition to linear transformations, contains reflections (transformations by reciprocal radii). In this case the existence of the function has been rigorously established in earlier work of Schwarz and Weierstrass, inasmuch as one prefers not to appeal to general Riemannian principles. See Volume 70 of Borchardt, Abbildung der Halbebene auf Kreisbogenpolygone.

7. Even in this special case I haven’t yet been able to find all the “groupes discontinus”; I have only established that there are many for which there is no determined fundamental circle, so that the analogy with non-Euclidean geometry (with which, by the way, I am very familiar) does not hold. If you take, for example, any polygon, the sides of which are any tangentially incident circles whatsoever, then generation via symmetry will always yield a groupe discontinu.

8. You will doubtless find answers to the other questions you pose in your letter in the articles I am sending you, in particular to those concerning the plural form of “Modulfunktionen” and, especially, “Fundamentalpolygonen”.

Hoping to hear from you again soon,

your very devoted

F. Klein

PTrL. Translated by S.A. Walter from the transcription in German by Nörlund (1923) (§ 4-47-3), in view of the translation by R. Burns in Saint-Gervais (2016). See also the French translation (§ 7-2-34).

Time-stamp: "27.04.2021 19:44"

Notes

• 1 Klein (1879b, a, c, 1876, 1877c, 1877a, 1877b).
• 2 Schwarz (1873).
• 3 Dedekind (1877).
• 4 Fuchs (1877), reed. Fuchs and Schlesinger (1906, 85–111).
• 5 Mapping the half-plane on polygons of circular arcs, by which title Klein presumably means to refer to Schwarz (1869); reed. Schwarz (1890, 65–83).