On the sum of the square roots of the first n natural numbers Journal of the Indian Mathematical Society, VII, 1915, 173 – 175 1. Let φ1 (n) = √ √ √ √ √ √ 1 + 2 + 3 + · · · + n − C1 + 23 n n + 12 n − 61 ν=∞ X ν=0 p p { (n + ν) + (n + ν + 1)}−3 , where C1 is a constant such that φ1 (1) = 0. Then we see that h i p p p φ1 (n) − φ1 (n + 1) = − (n + 1) + 23 (n + 1) (n + 1) + 12 (n + 1) p √ √ √ −( 32 n n + 12 n) + 16 { n − (n + 1)}3 = 0. But φ1 (1) = 0. Hence φ1 (n) = 0 for all values of n. That is to say n o−3 p √ √ √ √ √ √ √ √ 2 1 1 1 + 2 + 3 + 4 + · · · + n = C1 + 3 n n + 2 n + 6 n + (n + 1) np o−3 np o−3 p p + (n + 1) + (n + 2) (n + 2) + (n + 3) + + · · · . (1) But it is known that 1 C1 = 4π 1 1 1 √ + √ + √ + ··· . 1 1 2 2 3 3 Putting n = 1 in (1) and using (2), we obtain 1 1 1 1 √ √ √ + √ + √ + √ + ··· 2π √ ( 1)3 ( 1 + 2)3 ( 2 + 3)3 ( 3 + 4)3 1 1 1 1 =3 √ + √ + √ + √ + ··· . ( 1)3 ( 2)3 ( 3)3 ( 4)3 2. Again let √ √ √ √ √ √ φ2 (n) = 1 1 + 2 2 . . . + n n − C2 + 25 n2 n + 12 n n + 81 n 1 − 40 ν=∞ X ν=0 p p [ (n + ν) + (n + ν + 1)]−5 , (2) (3) Paper 9 60 where C2 is a constant such that φ2 (1) = 0. Then we have p φ2 (n) − φ2 (n + 1) = −(n + 1) (n + 1) p p p +{ 52 (n + 1)2 (n + 1) + 21 (n + 1) (n + 1) + 81 (n + 1)} p √ √ √ √ 1 { n − (n + 1)}5 = 0. −{ 25 n2 n + 12 n n + 18 n} + 40 But φ2 (1) = 0. Hence φ2 (n) = 0. In other words √ √ √ √ √ √ √ 1 1 + 2 2 + 3 3 + · · · n n = C2 + 52 n2 n + 21 n n + 18 n + But it is known that h √ p { n + (n + 1)}−5 i p p p p +{ (n + 1) + (n + 2)}−5 + { (n + 2) + (n + 3)}−5 + · · · . (4) 3 C2 = − 16π 2 1 40 1 1 1 √ + √ + √ + ··· . 12 1 22 2 32 3 It is easy to see from (4) and (5) that 1 1 1 1 2 √ √ √ √ 2π + √ + √ + √ + ··· ( 1)5 ( 1 + 2)5 ( 2 + 3)5 ( 3 + 4)5 1 1 1 1 = 15 √ + √ + √ + √ + ··· . ( 1)5 ( 2)5 ( 3)5 ( 4)5 (5) (6) 3. The corresponding results for higher powers are not so neat as the previous ones. Thus for example √ √ √ √ √ 5 n) 12 1 + 22 2 + 32 3 + · · · + n2 n = C3 + n( 27 n3 + 12 n2 + 24 p p p √ 1 − 96 [{ n + (n + 1)}−3 + { (n + 1) + (n + 2)}−3 + · · ·] h √ p p p 1 + 224 { n + (n + 1)}−7 + { (n + 1) + (n + 2)}−7 i p p +{ (n + 2) + (n + 3)}−7 + · · · ; and so on. √ √ √ √ 7 2 7 13 1 + 23 2 + · · · + n3 n = C4 + n( 92 n4 + 12 n3 + 24 n − 384 ) h √ i p p p 1 { n + (n + 1)}−5 + { (n + 1) + (n + 2)}−5 + · · · − 192 p p p √ 1 [{ n + (n + 1)}−9 + { (n + 1) + (n + 2)}−9 + · · ·]; + 1152 (7) (8) On the sum of the square roots of the first n natural numbers 61 The constants C3 , C4 , . . . can be ascertained from the well-known result that the constant in the approximate summation of the series 1r−1 + 2r−1 + 3r−1 + · · · + nr−1 is 1 1 1 2Γ(r) 1 + r + r + r + · · · cos 12 πr, (9) (2π)r 1r 2 3 4 provided that the real part of r is greater than 1. 4. Similarly we can shew, by induction, that √ 1 1 1 1 1 √ + √ + √ + · · · + √ = C0 + 2 n + √ n 2 n 1 2 3 ) ( √ p p p −3 −3 } { { n + (n + 1)} (n + 1) + (n + 2) p p + + ··· , − 12 {n(n + 1)} {(n + 1)(n + 2)} The value of C0 can be determined as follows: from (10) we have p 1 1 1 1 √ + √ + √ + ··· + p − 2 (2n) → C0 , 1 2 3 (2n) as n → ∞. Also 2 1 1 1 1 √ + √ + √ + ··· + p 2 4 6 (2n) as n → ∞. Now subtracting (12) from (11) we see that ! p √ − 2 (2n) → C0 2, (10) (11) (12) √ 1 1 1 1 1 √ − √ + √ − √ + ··· − p → C0 (1 − 2), as n → ∞. 1 2 3 4 (2n) That is to say C0 = −(1 + √ 2) 1 1 1 1 √ − √ + √ − √ + ··· . 1 2 3 4 We can also shew, by induction, that √ √ √ √ √ √ 1 + 2 + 3 + · · · + n = C1 + 32 n n + 12 n + (13) 1 √ 24 n # " √ p p p { n + (n + 1)}−5 { (n + 1) + (n + 2)}−5 p p + + · · · . (14) {n(n + 1)} {(n + 1)(n + 2)} √ √ √ √ The asymptotic expansion of 1 + 2 + 3 + · · · + n for large values of n can be shewn to be √ √ 1 1 1 1 2 1 − + − ··· , (15) C1 + 3 n n + 2 n + √ n 24 1920n2 9216n4 1 − 24 by using the Euler-Maclaurin sum formula.