A formula for the number of Latin squares (Q1208376)

The authors establish an explicit formula for the number of Latin squares of order \(n\), namely \[ L_ n=n!\sum_{A\in B_ n}(-1)^{\sigma_ 0(A)}{\text{per} A\choose n}, \] where \(B_ n\) is the set of \((0,1)\) matrices of size \(n\times n\), \(\sigma_ 0(A)\) is the number of zeros of the matrix \(A\) and per \(A\) denotes the permanent of the matrix \(A\). The authors claim that \[ L_ n=\sum_{A\in B_ n}(-1)^{\sigma_ 0(A)}(\text{per} A)^ n \] also holds. One can find in this paper explicit formulas for the number of Latin rectangles of size \(k\times n\), too. The latter formulas are trivial consequences of the formulas for the quantity \(L_ n\). Reviewer's remarks: (1) The authors use \textit{H. J. Ryser}'s book [Combinatorial mathematics. The Carus Mathematical Monographs, No. 14. Published by The Mathematical Association of America. New York: John Wiley and Sons (1963; Zbl 0112.24806)] as the only reference and seem to be unaware of the origin of their formulas and also what happened in the last 30 years in this field. (2) The authors claim that the values of \(L_ n\) are known for \(n=2,3,4,5,6\), and \(7\). \(L_ 8\) and \(L_ 9\) are also known (see \textit{M. B. Wells} [J. Comb. Theory 3, 98--99 (1967; Zbl 0166.01001)], \textit{J. W. Brown} [J. Comb. Theory 5, 177--184 (1968; Zbl 0165.02901)], \textit{S. E. Bammel} and \textit{J. Rothstein} [Discrete Math. 11, 93--95 (1975; Zbl 0304.05007)], \textit{V. L. Arlazarov}, \textit{A. M. Baraev}, \textit{Ya. Yu. Gol'fand} and \textit{I. A. Faradzhev} [Algorithmic studies in combinatorics, Work Collect., Moskva 1978, 129--141 (1978; Zbl 0483.05017)]). All the known values of \(L_ n\) can be found on page 144 of the reviewer's joint book with \textit{A. D. Keedwell} [Latin squares and their applications (1974; Zbl 0283.05014)]. The latter values were confirmed in a very recent paper of \textit{G. L. Mullen} and \textit{D. Purdy} [J. Comb. Math. Comb. Comput. 13, 161--165 (1993; Zbl 0751.05015)]. (3) The authors refer to a lower bound of \(L_ n\) given in Ryser's book mentioned above. They completely ignore that in the meantime van der Waerden's conjecture on permanents has been solved (see \textit{G. P. Egorichev}'s preprint Krasnoyarsk 1980 and \textit{D. I. Falikman} [Math. Notes 29, 475--479 (1981); translation from Mat. Zametki 29, 931--938 (1981; Zbl 0475.15007)]. If one combined the results above with the formulas of the authors, one could obtain a new and very good lower bound for \(L_ n\). (\textit{H. J. Ryser}'s posthumous joint book with \textit{R. A. Brualdi} [Combinatorial matrix theory. Cambridge (1991; Zbl 0746.05002)] contains some recent results on the connections between permanents and Latin squares.) (4) The basic ideas which led to the explicit formula for \(L_ n\), originated from \textit{P. A. MacMahon} [Trans. Camb. Phil. Soc. 16, 262--290 (1898; JFM 30.0227.11) and Nature 65, 447--452 (1901; JFM 33.0232.04)]. (5) MacMahon's ideas were used recently in two joint papers by the reviewer and \textit{A. D. Keedwell} [Ars Comb. 25A, 109--126 (1988; Zbl 0644.05014) and Util. Math. 34, 73--83 (1988; Zbl 0683.05012)]. The interested reader can find further recent results on the interpretation of MacMahon's ideas in a joint paper of the reviewer, \textit{A. D. Keedwell} and \textit{G. L. Mullen} [Ars Comb. 28, 201--202 (1989; Zbl 0701.05044)] and in a joint paper of the reviewer and \textit{G. L. Mullen} [Discrete Math. 111, 157--163 (1993; Zbl 0789.05007)].