On sums of a prime and four prime squares in short intervals (Q5920718)

The authors consider the problem of representing integers as sums of four squares of primes lying in short intervals. It is proved that for large \(N\) and \(U = N^{\frac {9}{20} + \varepsilon}\), all but at most \(O(U^{1-\varepsilon})\) integers \(n\equiv 4 \pmod{24}\) in the interval \([\frac 45 N , \;\frac 45 N + U]\) are representable as sums of four squares of primes, each lying in the interval \([\sqrt{\frac N5} - U, \;\sqrt{\frac N5} + U]\). The size of \(U\) here represents an improvement over previous results, and this follows as an application of an estimate for exponential sums over the square of primes in short intervals, which \textit{J. Liu, G. Liu} and \textit{T. Zhan} obtained recently [Sci. China, Ser. A 49, No. 5, 611--619 (2006; Zbl 1106.11033)]. The authors then use this result to deduce that each large integer \(N\not\equiv 1 \pmod 3\) can be represented as \(N = p + p^2_1 + \cdots + p^2_4\) where \(p, p_1, \dots, p_4\) are primes and \(|p - \frac N5 |\leq U\). But the short interval for \(p\) here seems to be beyond our present technology.