594 PI MU EPSILON JOURNAL PROBLEMS AND SOLUTIONS 595 c) Using the same technique as in Part (b), we see that the first k terms vanish and we have II L (i - k)(i - k + 1) ..·(i)(i + 1)(i + 2)··· i=l n-k = L (i)(i + 1)(i + 2)···
596 PI MU EPSILON JOURNAL PROBLEMS AND SOLUTIONS 597 Table 2. Solutions to the Generalized Problem. X; Since I 0 I can be written as a sum of two squares in exactly one way, I 0 I = <strong>10</strong>0 + I, and since 2jy- sj is even, then we must have 2jy- sj := IO and hence y = 0 or y = I 0. Therefore, there are no integer solutions with I:$;; y ~ 9. 2 3 4 8c + I 8c + 3 8c + 2 64c 2 + 24c + I 64c 2 + 40c + 5 64c 2 + 32c + 3 5I2c 3 + 320c 2 5I2c 3 + 448c 2 5I2c 3 + 384c 2 + 48c + I + II2c + 7 + 80c + 4 Also solved by Paul S. Bruckman, Highwood, /L, Russell Euler and Jawad Sadek, <strong>No</strong>rthwest Missouri State University, Maryville, Mark Evans, Louisville, KY, Richard I. Hess, Rancho Palos Verdes, CA, <strong>Mu</strong>rray S. Klamkin, University of Alberta, Canada, William H. Peirce, Delray Beach, FL. H.-J. Seiffert, Berlin, Germany, and the Proposer. Klamkin referred to his article "Perfect Squares of the form (m 2 - I)~ + 1," Math. Mag., 40(I969)III-113. 898. [Fall 1996] Proposed by PaulS. Bruckman, Edmonds, Washington. An n-digit number N is defined to be a base <strong>10</strong> Armstrong number of order n if n-1 n-1 N = E dklok = E dk", k=O k=O where the dk are decimal digits, with dn-J > 0. (See Miller and Whalen, "Armstrong Numbers: I 53= I 3 +5 3 + 3 3 ," Fibonacci Quarterly30.3, (I992), pp. 22I-224.) Prove that there are no base ten Armstrong numbers of order 2; that is, prove the impossibility of the equation I Oy + x = x 2 + y, where x and y are integers with 0 :$;; x :$;; 9 and I :$;; y :$;; 9. I. Solution by Charles R. Diminnie, San Angelo, Texas. By completing the squares, we reduce the equation to (2x- I) 2 + 4(y- 5) 2 = IOI. II. Solution by Bob Prielipp, University of Wisconsin-Oshkosh, Oshkosh, Wisconsin. Since I Oy - x(x - 1) = y, then y is even, so y is even. Thus y ~ 2, 4, 6, or 8. If y = 2 or 8, then the given equation becomes r =X+ I6, that is, (2x- I) 2 = 65. When x = 4 or 6, the equation becomes x 2 = x + 24, that is, (2x- I) 2 = 97. Since neither 65 nor 97 is a square number, there is no solution with the given constraints. III. Solution by James Campbell, University of Missouri, Columbia, Missouri. We must have (<strong>10</strong>- y)y = x(x- I). <strong>No</strong>w 9 S (<strong>10</strong>- y)y ~ 25 whenever 1 :$;; y :$;; 9 and x(x - 1) is an increasing function, so if there is a solution, x = 4 or x = 5. Substituting these values in for x, we get the equations y - 1 Oy + 12 = 0 and y - 1 Oy + 20 = 0. In neither case is the discriminant a perfect square, so no integer solution exists. IV. Solution by Robert C. Gebhardt, Hopatcong, New Jersey. By the quadratic formula, y = 5 ± V25 - x 2 + X. The radicand, 25 - x 2 + x, is a perfect square only if x = 0 or x = I. Therr y = 0 or y = <strong>10</strong>, neither of which is permitted. Thus there is no solution.