Values of Ramanujan's signature hypergeometric functions $F_r(x)$

edit · history · discussion · files · long url · special functions elliptic curves special values
Numbers
$r$
$x$ 
$F_r(x)$
2
1/2:
1.180340599016096226045337940558488587233716634881447299515864399404304180720715794978458616195807954
comment: A Gauss second-summation value, equal to $2/B((r+1)/(2r),(2r-1)/(2r))$.
2
0:
1
2
1/4:
1.073182007149364375052841707970349795669526818631368625787232257139788343028625222768722717743272285
2
3/4:
1.372880500618350164697637575007806058094538625348940055855540038870350633240892464904292347403572711
2
1/3:
1.103845772797212254324837694207631411924760802239391086997330526905118375631119575423168443314817465
2
2/3:
1.291675482135082644544318504760649923199863516052326580950539239585172786733440152066839813967604411
2
1/9:
1.029660375463762623523228181255854543570903969011705041891655029152762974930669348678219260552679533
2
8/9:
1.609773010724046562579262561955524693504290227947052938680848385709682514542937834153084076614908428
3
1/2:
1.159595266963928365769992051570020881945165263439782855263105059747973757205286258658085251051146071
comment: A Gauss second-summation value, equal to $2/B((r+1)/(2r),(2r-1)/(2r))$.
3
0:
1
3
1/4:
1.064925231128516498437819437604916718248282322874928859475392443772515623080612622146534197451821389
3
3/4:
1.328778200157056363878073355047905281137317875982473604891262407713582251790491637816553512224526597
3
1/3:
1.092060148081664394316757805253605515050933694436801793669960490099214146997645089823341657245948803
3
2/3:
1.257546863411625467745864646361554193208060216425676017110925090768116861297340574645820318884594719
3
1/9:
1.026343520841566885711818335651046428274180826014178666730958109479707060905841596720020526615780503
3
8/9:
1.535812809958194137289311853725147146475663454326775827621496728753103561572492083134212409167445065
4
1/2:
1.133915559726082732440156510347438791344295356805403611236317061400497913023777214458207113170347034
comment: A Gauss second-summation value, equal to $2/B((r+1)/(2r),(2r-1)/(2r))$.
4
0:
1
4
1/4:
1.054648614831467047906718149577441840630273364763410414893082861937658163667864603886693740230931228
4
3/4:
1.274611308136601952102809952021133623397842247250675660982363975799881719700368808191505969508039952
4
1/3:
1.077415715750701203246257098513742012348092021182187093026426589300763704600445579059592620512413547
4
2/3:
1.215502554868454386941982217997043941288277531484048915816265650173538426379192234806255033882957850
4
1/9:
1.022204943866080919860161806857062989134020137789994880270437701856939901361339037391092749731504665
4
8/9:
1.445616095140239952072718914467638749495268402565399677644614166992051212768192453714749044140042691
6
1/2:
1.098430696839862068942935161608698822283938481692253723027033945469518521905843821564973446230505442
comment: A Gauss second-summation value, equal to $2/B((r+1)/(2r),(2r-1)/(2r))$.
6
0:
1
6
1/4:
1.040343865043574145416878113346879640594465685268916549682386133655136588848837323908014059028010148
6
3/4:
1.200547022551705970623606133782633712417243715140519123069736520860927175815592908672157063181494441
6
1/3:
1.057076261613892128402122962965324995470501889069986725795375244685741777798779952700216143078965700
6
2/3:
1.157778216364893874950561674893505207806365917352306655090978043246823064194951989915747252390826364
6
1/9:
1.016424899916147157209316669882466843802873309665909634999036140962440390965478130838565284637450640
6
8/9:
1.323467526740777845709483066819782686939481648265356144701869051624955722200944233060569225543086875
Definition
This table gives the real values $F_r(x)={}_2F_1(1/r,1-1/r;1;x)$ of Gauss's hypergeometric function [3] for signatures $r\in\{2,3,4,6\}$ and rational arguments $0\leq x<1$. The argument $x$ is Gauss's hypergeometric variable.
Parameters
$r$
—   signature ($r\in\{2,3,4,6\}$)
$x$
—   hypergeometric argument ($0\leq x<1$)
Formulas
(1)
With $(a)_n$ denoting the Pochhammer symbol, $F_r(x)=\sum_{n=0}^{\infty} ((1/r)_n(1-1/r)_n/(n!)^2)x^n$ for $|x|<1$ [3].
(2)
For $y=1-x$ and $0<x<1$, $F_r(x)=\frac{\sin(\pi/r)}{\pi}\sum_{n=0}^{\infty} \frac{(1/r)_n(1-1/r)_n}{(n!)^2} \bigl(2\psi(n+1)-\psi(n+1/r)-\psi(n+1-1/r)-\log y\bigr)y^n$, where $\psi$ is the digamma function. This is the $c=a+b$ case of the continuation formula [5].
(3)
As $x\to1^-$, $F_r(x)/(-\log(1-x))\to\sin(\pi/r)/\pi$.
(4)
$F_2(x)=2K(x)/\pi$, where $K$ is the complete elliptic integral of the first kind with parameter $m=x$.
(5)
$F_2(1-b^2)=1/\operatorname{AGM}(1,b)$. The cubic analogue is $F_3(1-b^3)=1/\operatorname{AGM}_3(1,b)$, where $\operatorname{AGM}_3$ is the common limit of $a_{n+1}=(a_n+2b_n)/3$ and $b_{n+1}=\sqrt[3]{b_n(a_n^2+a_nb_n+b_n^2)/3}$ [2].
(6)
$F_4(x)=\sqrt{1+z}\,F_2(z)$ with $z=(1-\sqrt{1-x})/(1+\sqrt{1-x})$ [5].
(7)
For $r\in\{2,3,4,6\}$, $F_r(1/2)=2/B((r+1)/(2r),(2r-1)/(2r))= \sqrt{\pi}/(\Gamma((r+1)/(2r))\Gamma((2r-1)/(2r)))$ [4], with $B(a,b)$ as in [6].
Comments
(8)
Signature $2$ is the classical elliptic-integral case, related to $K(m)$ by (4). Signatures $3$, $4$ and $6$ are the alternative-base theories of Ramanujan developed by Berndt, Bhargava and Garvan [1]; in these four cases the two upper hypergeometric parameters are $a$ and $1-a$, with $a=1/r$.
(9)
$F_r(0)=1$, and $F_r(x)$ diverges logarithmically as $x\to1^-$ by (3).
(10)
At $x=1/2$, Gauss's second summation theorem gives the Gamma and Beta quotients in (7). The signature-$2$ entry agrees with [7]. The OEIS entry [8] writes the signature-$3$ value as $2^{-2/3}/L$, with $L=\Gamma(1/3)\Gamma(5/6)/\Gamma(1/6)$.
(11)
The signatures $r=2,3,4,6$ retain their alternative-base elliptic meaning. The half argument has the gamma quotient from Gauss's second summation theorem [10]. Complementary pairs $(x,1-x)$ with $x=1/4,1/3,1/9$ provide a bounded low-complexity reference set on both sides of the complementary centre; this list is our exploratory choice and is not asserted to consist of singular moduli. The origin supplies the exact hypergeometric normalization.
Programs
(P1)
Python
from balanced_values import BalancedValues
from numberdb._write import to_text
reference = BalancedValues()
for parameters in reference.enumerate():
    print(parameters, to_text(reference.value(parameters, reference.digits), reference.digits))
References
[1]
Bruce C. Berndt, S. Bhargava and Frank G. Garvan, Ramanujan's theories of elliptic functions to alternative bases, Trans. Amer. Math. Soc. 347 (1995), no. 11, 4163-4244. (doi) (zbMATH) (MR)
[2]
J. M. Borwein and P. B. Borwein, A cubic counterpart of Jacobi's identity and the AGM, Trans. Amer. Math. Soc. 323 (1991), no. 2, 691-701. (doi) (MR)
Links
Similar tables
Complete elliptic integral of the first kind $K(m)$ —   signature $2$ is related to $K(m)$ by (4)
Values of the arithmetic-geometric mean —   signature $2$ has the AGM expression and signature $3$ has the cubic analogue in (5)
Values of the Beta function $B(a,b)$ at pairs of rational numbers —   the entries with $x=1/2$ are reciprocal Beta values by (7)
Values of the Gamma function at rational numbers —   the entries with $x=1/2$ are Gamma quotients by (7)
Data properties
Entries are of type: real number
How they were obtained:

The four values at $x=0$ are exactly $F_r(0)=1$. The other 28 values are evaluated from exact rational inputs with Arb's hypergeometric function at 100 decimal digits plus 160 working guard bits. Each stored decimal interval contains the corresponding finite enclosure; the minimum measured relative accuracy is 482 bits.

more

Independent mpmath evaluations at 150 and 200 decimal digits check all 32 entries, and the four values at $x=1/2$ are also checked against the Gamma quotient in (7). Exact-origin, nonzero, shifted-last-place rejection and enumeration controls check the validation itself. The attached balanced-evidence.json records these checks, and balanced_checks.py reproduces them. Controls for other functions in that shared code are not additional checks of this table.

The current generate.py and balanced-selection.json reproduce the full selection without changing the retained numerical strings or row annotations. Older auxiliary attachments remain historical sources, not the default enumeration.

Table is complete: no (it holds $F_r(x)$ for each signature $r=2,3,4,6$ at $x=0,1/2,1/4,3/4,1/3,2/3,1/9,8/9$; the complementary pairs are exploratory references, not a catalogue of arithmetic moduli, and the origin is exact)