Given the squared nome $q=e^{2i\pi\tau}$ with $|q|\lt1$, define


Question: How do we show that $$H(q)\overset{\color{red}{?}}=\frac{2\,q^{1/2}\,\vartheta_3(0,q^2)}{\vartheta_2(0,q^2)}$$ where $\vartheta_n(0,q)$ are jacobi theta functions

  • $\begingroup$ Try converting it to a sum using euler formula $\endgroup$ – Oussama Boussif Sep 11 '15 at 17:32
  • $\begingroup$ @Oussama Boussif:unfortunately it does not have the simple form of euler's continued fraction formula $\endgroup$ – Nicco Sep 11 '15 at 17:39
  • $\begingroup$ Is there any sequence defining $2(1+q^2),(1+q)(1+q^3),...$? $\endgroup$ – Oussama Boussif Sep 11 '15 at 18:02
  • 1
    $\begingroup$ The pattern of the cfrac starts developing from $$2q^2(1+q^4)$$ and goes like $$q^3(1+q)(1+q^5),q^4(1+q^2)(1+q^6),q^5(1+q^3)(1+q^7)$$ so on,where the pattern is clear $\endgroup$ – Nicco Sep 11 '15 at 18:30
  • $\begingroup$ @OussamaBoussif: My answer below shows the pattern even prior to Nicco's $2q^2(1+q^4)$. $\endgroup$ – Tito Piezas III Sep 12 '15 at 13:56

In a related answer, I used a formula by Ramanujan, proved by Adiga et al. (1985): $$\small\frac{(-a;q)_\infty\,(-b;q)_\infty - (a;q)_\infty\,(b;q)_\infty} {(-a;q)_\infty\,(-b;q)_\infty + (a;q)_\infty\,(b;q)_\infty} = \cfrac{a+b}{1-q+\cfrac{(a+bq)(aq+b)}{1-q^3+\cfrac{q\,(a+bq^2)(aq^2+b)} {1-q^5+\cfrac{q^2(a+bq^3)(aq^3+b)}{1-q^7+\cdots}}}}\tag{*}$$ Applying it here can be done by setting $a=q_2^3$, $b=q_2^{-1}$ where $q_n=\exp(2\pi\mathrm{i}\tau/n)$, so $q_n^n=q$. Thus we get $$\begin{align} \frac{H(q)}{2q_2} &= q_2^{-1}\,\cfrac{1+q^2}{1-q+\cfrac{(1+q)(1+q^3)}{1-q^3+\cfrac{2q^2(1+q^4)} {1-q^5+\cfrac{q^3(1+q)(1+q^5)}{1-q^7+\cfrac{q^4(1+q^2)(1+q^6)} {1-q^9+\cdots}}}}} \\ &= \frac{(-q_2^3;q)_\infty\,(-q_2^{-1};q)_\infty - (q_2^3;q)_\infty\,(q_2^{-1};q)_\infty} {(-q_2^3;q)_\infty\,(-q_2^{-1};q)_\infty + (q_2^3;q)_\infty\,(q_2^{-1};q)_\infty} \\ &= \frac{(-q_2^3;q)_\infty^2\,(-q_2^{-1};q)_2 - (q_2^3;q)_\infty^2\,(q_2^{-1};q)_2} {(-q_2^3;q)_\infty^2\,\underbrace{(-q_2^{-1};q)_2}_{(1+q_2)^2/q_2} + (q_2^3;q)_\infty^2\,\underbrace{(q_2^{-1};q)_2}_{-(1-q_2)^2/q_2}} \\ &= \frac{(-q_2;q)_\infty^2 + (q_2;q)_\infty^2} {(-q_2;q)_\infty^2 - (q_2;q)_\infty^2} \\ &\stackrel{[1]}{=} \frac{\vartheta_3(0,q_2) + \vartheta_3(0,-q_2)} {\vartheta_3(0,q_2) - \vartheta_3(0,-q_2)} \\ &\stackrel{[2]}{=} \frac{\vartheta_3(0,q^2)}{\vartheta_2(0,q^2)} \end{align}$$ where [1] follows from the product representation $$\begin{align} \vartheta_3(0,q_2) &= (-q_2;q)_\infty^2\,(q;q)_\infty \end{align}$$ and [2] follows from splitting the series representation of $\vartheta_3$ into odd and even parts, resulting in: $$\begin{align} \vartheta_3(0,q_2) + \vartheta_3(0,-q_2) &= 2\vartheta_3(0,q^2) \\ \vartheta_3(0,q_2) - \vartheta_3(0,-q_2) &= 2\vartheta_2(0,q^2) \end{align}$$ More details and background on the Ramanujan formula can be found at the other answer.


(A partial answer.) Compare the three cfracs of similar form,


$$U(-q)\; =\; 1\;=\;\small\cfrac{(1+q^1)}{1-q+\cfrac{q(1+q^0)(1+q^2)}{1-q^3+\cfrac{q^2(1+q)(1+q^3)}{1-q^5+\cfrac{q^3(1+q^2)(1+q^4)}{1-q^7+\cfrac{q^4(1+q^3)(1+q^5)}{1-q^{9}+\ddots}}}}}\tag2$$

$$S(q) = \frac{1}{q^{1/2}}\frac{\vartheta_2(0,q^2)}{\vartheta_3(0,q^2)}=\small\cfrac{(1+q^0)}{1-q+\cfrac{q(1+q)(1+q)}{1-q^3+\cfrac{q^2(1+q^2)(1+q^2)} {1-q^5+\cfrac{q^3(1+q^3)(1+q^3)}{1-q^7+\cfrac{q^4(1+q^4)(1+q^4)} {1-q^9+\ddots}}}}}\tag3$$

where the pattern of one cfrac to the next is clear. We have $(1)$ as your proposed equality, $(2)$ is the one in this post (with the small change $q \to -q$ for aesthetics), and $(3)$ was established by Somos in A079006. One can then see the nice fact that,

$$H(q)\,U(-q)\,S(q) = 1$$

In fact, $H(q)$ and $S(q)$ are reciprocals. (This is the second reciprocal pair you have found after this, so I assume you are using a general method?)

P.S. Since $(2)$ belongs to an infinite family involving ratios of form $(1-aq^n)$, then $(1),(3)$ may have an infinite family as well.

  • $\begingroup$ @ Tito Piezas:yes I am using a general method,which yields infinitely many cfracs of interest,however there is a difference between a rigorous proof and method,the former which is what I'm least good at due to lack of formal training. $\endgroup$ – Nicco Sep 12 '15 at 3:50
  • $\begingroup$ @Nicco: Your cfracs look very nice and aesthetic. But they would be nicer if, instead of being isolated results, they would be placed in the context of a general method/theory. $\endgroup$ – Tito Piezas III Sep 12 '15 at 13:29
  • $\begingroup$ @Nicco: Take a look at my version of $H(q)$. The pattern is clearer. :) $\endgroup$ – Tito Piezas III Sep 12 '15 at 13:44
  • $\begingroup$ @ Tito PiezasIII:wow,very impressive indeed,your version is revealing. $\endgroup$ – Nicco Sep 12 '15 at 13:51
  • $\begingroup$ @Nicco: Maybe you can use that trick of negative powers in your other cfracs. $\endgroup$ – Tito Piezas III Sep 12 '15 at 13:54

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.