A note on the functional equation \(f^{n}+g^{n}+h^{n}\)=1 and some complex differential equations (Q1409776)

The Fermat problem for meromorphic functions (in the complex plane) asks for a triple \((f,g,h)\) of such functions satisfying (*) \(f^n+g^n+h^n=1\), given \(n\geq2\). Their existence is wellknown for \(n=2,3,4\), while the cases \(n=5\), \(n=6\) have been recently solved by \textit{G. Gundersen} [Complex Variables, Theory Appl. 43, 293--298 (2001; Zbl 1022.30029), Analysis 18, 285--290 (1998; Zbl 0916.30028)]. On the other hand, \textit{W. Hayman} proved [Bayer. Akad. Wiss. Math.-Natur. Kl. Sitzungsber., 1--13 (1985; Zbl 0567.30002)] that no such triples exist for \(n\geq9\). The main result in this paper, for \(n\geq7\), assumes that \((f,g,h)\) satisfy (*) and that \(f,g,h\) are linearly independent, resulting in \[ \frac{n-6}nT^*(r)\leq \frac6{n-2}N\left(r,\frac1\Delta\right)+S^*(r), \] where \(T^*(r):=T(r,f^n)+T(r,g^n)+T(r,h^n)\), \(\Delta(z):=W(f^n,g^n,h^n)/(f^ng^nh^n)\), and \(S^*(r)\) is small (in the usual Nevanlinna theory sense) with respect to \(T^*(r)\). Two consequences of this result are: (a) Another proof for the nonexistence of \((f,g,h)\) for \(n\geq9\). (b) If \(n=8\), then \(W(f^8,g^8,h^8)=a(z)(f(z)g(z)h(z))^6\), where \(T(r,a(z))\) is small w.r.\ to \(T^*(r)\). In the special case of an entire triple \((f,g,h)\), the non-existence result for \(n\geq7\), due to Hayman, will be reproved. For \(n=6\), it follows for an entire triple \((f,g,h)\) that \(W(f^6,g^6,h^6)=b(z)(f(z)g(z)h(z))^4\), \(b(z)\) small. The proofs in this paper apply the Nevanlinna theory. The most important ingredient of the proofs is a detailed, careful analysis of the poles of \(f^n\), \(g^n\) and \(h^n\).