Using Raabe's test, examine the convergence of: sum
Using Raabe's test, examine the convergence of: $$ \sum_{n=1}^{\infty} \frac{\alpha(\alpha+1) \cdots(\alpha+n-1)\,\beta(\beta+1) \cdots(\beta+n-1)}{n!\,\gamma(\gamma+1) \cdots(\gamma+n-1)}. $$
Explanation of Method
Writing the term with the Pochhammer (rising-factorial) symbol $(\alpha)_n=\alpha(\alpha+1)\cdots(\alpha+n-1)$, the general term is $u_n=\frac{(\alpha)_n(\beta)_n}{n!(\gamma)_n}$ — this is the classical hypergeometric series ${}_2F_1(\alpha,\beta;\gamma;1)$. The ratio $u_n/u_{n+1}$ simplifies to a rational function of $n$, and Raabe's limit reduces neatly to $\gamma-\alpha-\beta+1$: Gauss's celebrated convergence criterion.
- Express $u_{n+1}$ via the Pochhammer recursions.
- Compute the ratio $u_n/u_{n+1}=\dfrac{(n+1)(\gamma+n)}{(\alpha+n)(\beta+n)}$.
- Evaluate the Raabe limit $r=\gamma-\alpha-\beta+1$ and resolve the boundary case.
Step 1 · Set up $u_n$ and $u_{n+1}$
With $u_n=\dfrac{(\alpha)_n(\beta)_n}{n!(\gamma)_n}$, using $(\alpha)_{n+1}=(\alpha)_n(\alpha+n)$ and likewise for $\beta,\gamma,n!$:
$$ u_{n+1} = \frac{(\alpha)_n(\alpha+n)(\beta)_n(\beta+n)}{n!(n+1)(\gamma)_n(\gamma+n)}. $$
Step 2 · Compute the ratio
$$ \frac{u_n}{u_{n+1}} = \frac{(n+1)(\gamma+n)}{(\alpha+n)(\beta+n)}. $$
Step 3 · Apply Raabe's test
Expanding both products: $(n+1)(\gamma+n)=n^2+n(1+\gamma)+\gamma$ and $(\alpha+n)(\beta+n)=n^2+n(\alpha+\beta)+\alpha\beta$, so:
$$ \frac{u_n}{u_{n+1}}-1 = \frac{n(\gamma-\alpha-\beta+1)+(\gamma-\alpha\beta)}{(\alpha+n)(\beta+n)}, \qquad r = \lim_{n\to\infty} n\left(\frac{u_n}{u_{n+1}}-1\right) = \gamma-\alpha-\beta+1. $$
So: converges if $\gamma>\alpha+\beta$; diverges if $\gamma<\alpha+\beta$; at $\gamma=\alpha+\beta$ ($r=1$), a finer test (Gauss's test) shows the series still diverges.
Summary Table
| Condition | Conclusion |
|---|---|
| $\gamma>\alpha+\beta$ | Converges |
| $\gamma=\alpha+\beta$ | Diverges (by Gauss's refinement) |
| $\gamma<\alpha+\beta$ | Diverges |
Additional Notes
- This is exactly Gauss's classical criterion for the hypergeometric series ${}_2F_1(\alpha,\beta;\gamma;x)$ evaluated at the boundary point $x=1$ of its unit radius of convergence, one of the most historically significant applications of Raabe-type tests.
- Problem 90's series $\sum\frac{(a)_n}{n!}\cdot\frac1{n^a}$ can be seen as a degenerate special case of this hypergeometric structure (formally $\beta\to\infty$ absorbed into the extra factor $n^{-a}$), which is why both problems land on Raabe limit $r=1$ or a similarly delicate boundary.
- The symmetry between $\alpha$ and $\beta$ in the convergence criterion ($\gamma>\alpha+\beta$, not favoring either) reflects the underlying symmetry of the hypergeometric function itself: ${}_2F_1(\alpha,\beta;\gamma;x)={}_2F_1(\beta,\alpha;\gamma;x)$.
Result
The hypergeometric series converges if and only if $\gamma>\alpha+\beta$: