enter image description here

I can't seem to find a way, squaring the expression would make more terms and would make it harder, I guess there must be something to do with the first and last terms as they sum to 100? or maybe difference of to squares, but i can't solve it.

  • $\begingroup$ Interesting. Crude estimates quickly tell you which of those answers is correct (all but one are too large). And a quick numerical calculation confirms the result, but I don't see a cheap algebraic mode of attack. $\endgroup$ – lulu Oct 18 '18 at 15:22
  • 5
    $\begingroup$ Numerical computation strongly supports the conjecture that $$\sum_{k=1}^{n^2-1}\sqrt{n+\sqrt{k}}=(1+\sqrt2)\sum_{k=1}^{n^2-1}\sqrt{n-\sqrt{k}}$$ for $n\geq2$ I don't have any idea how to attack it yet. $\endgroup$ – saulspatz Oct 18 '18 at 16:04
  • $\begingroup$ What is the source of this problem? $\endgroup$ – saulspatz Oct 18 '18 at 16:38

Useful fact:

$$ \sqrt{a\pm \sqrt{b}}= \sqrt{\frac{a+\sqrt{a^2-b}}{2}} \pm \sqrt{\frac{a-\sqrt{a^2-b}}{2}} \tag{1}$$


\begin{align} RHS^2 &= \frac{a+\sqrt{a^2-b}}{2}+\frac{a-\sqrt{a^2-b}}{2} \pm 2 \sqrt{\left( \frac{a+\sqrt{a^2-b}}{2} \right) \left( \frac{a-\sqrt{a^2-b}}{2} \right)} \\ &= a \pm 2\sqrt{\frac{a^2-(a^2-b)}{4}} \\ &= a \pm 2\sqrt{\frac{b}{4}} \\ &= a \pm \sqrt{b} \\ &= LHS^2 \\ LHS &= RHS \qquad (a^2 \ge b \ge 0) \end{align}


$$S_{\pm}=\sum_{j=1}^{n^2-1} \sqrt{n \pm \sqrt{j}} \tag{2}$$

and reverse the order of summation by taking $\, k=n^2-j$, then

$$S_{\pm}=\sum_{k=1}^{n^2-1} \sqrt{n \pm \sqrt{n^2-k}} \tag{3}$$

Now by $(1)$, $(2)$ and $(3)$,

$$S_{\pm}=\frac{S_+ \pm S_-}{\sqrt{2}} \implies \frac{S_+}{S_-}=1+\sqrt{2}$$

  • 2
    $\begingroup$ sharp answer indeed ! (+1) $\endgroup$ – G Cab Oct 18 '18 at 17:58
  • $\begingroup$ Im really sorry but could you expound on how you got the first eq $\endgroup$ – SuperMage1 Oct 18 '18 at 21:32
  • $\begingroup$ @Supermage1 Well that could be easy. Just form perfect squares inside the radicals. So just solve $xy=b/4$ and $x+y=a$ to get $x, y$ ans hence the the original radical would equal $\sqrt x+\sqrt y$ $\endgroup$ – Rohan Shinde Oct 19 '18 at 2:26
  • $\begingroup$ @SuperMage1 I've added a proof for the identity $(1)$ in my answer. And what Manthanein suggested is to find $x$ and $y$ such that $LHS=\sqrt{x+y \pm 2\sqrt{xy} \,}=|\sqrt{x} \pm \sqrt{y} \,|$. $\endgroup$ – Ng Chung Tak Oct 19 '18 at 9:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.