s21bc returns a value of the symmetrised elliptic integral of the second kind.

# Syntax

C# |
---|

public static double s21bc( double x, double y, double z, out int ifail ) |

Visual Basic |
---|

Public Shared Function s21bc ( _ x As Double, _ y As Double, _ z As Double, _ <OutAttribute> ByRef ifail As Integer _ ) As Double |

Visual C++ |
---|

public: static double s21bc( double x, double y, double z, [OutAttribute] int% ifail ) |

F# |
---|

static member s21bc : x : float * y : float * z : float * ifail : int byref -> float |

#### Parameters

- x
- Type: System..::..Double
*On entry*: the arguments $x$, $y$ and $z$ of the function.

- y
- Type: System..::..Double
*On entry*: the arguments $x$, $y$ and $z$ of the function.

- z
- Type: System..::..Double
*On entry*: the arguments $x$, $y$ and $z$ of the function.

- ifail
- Type: System..::..Int32%
*On exit*: ${\mathbf{ifail}}={0}$ unless the method detects an error or a warning has been flagged (see [Error Indicators and Warnings]).

#### Return Value

s21bc returns a value of the symmetrised elliptic integral of the second kind.

# Description

s21bc calculates an approximate value for the integral

where $x$, $y\ge 0$, at most one of $x$ and $y$ is zero, and $z>0$.

$${R}_{D}\left(x,y,z\right)=\frac{3}{2}\underset{0}{\overset{\infty}{\int}}\frac{dt}{\sqrt{\left(t+x\right)\left(t+y\right){\left(t+z\right)}^{3}}}$$ |

The basic algorithm, which is due to Carlson (1979) and Carlson (1988), is to reduce the arguments recursively towards their mean by the rule:

For $n$ sufficiently large,

and the function may be approximated adequately by a fifth order power series

where ${S}_{n}^{\left(m\right)}=\left({X}_{n}^{m}+{Y}_{n}^{m}+3{Z}_{n}^{m}\right)/2m\text{.}$ The truncation error in this expansion is bounded by $\frac{3{\epsilon}_{n}^{6}}{\sqrt{{\left(1-{\epsilon}_{n}\right)}^{3}}}$ and the recursive process is terminated when this quantity is negligible compared with the machine precision.

$$\begin{array}{lcl}{x}_{0}& =& x,{y}_{0}=y,{z}_{0}=z\\ {\mu}_{n}& =& \left({x}_{n}+{y}_{n}+3{z}_{n}\right)/5\\ {X}_{n}& =& \left(1-{x}_{n}\right)/{\mu}_{n}\\ {Y}_{n}& =& \left(1-{y}_{n}\right)/{\mu}_{n}\\ {Z}_{n}& =& \left(1-{z}_{n}\right)/{\mu}_{n}\\ {\lambda}_{n}& =& \sqrt{{x}_{n}{y}_{n}}+\sqrt{{y}_{n}{z}_{n}}+\sqrt{{z}_{n}{x}_{n}}\\ {x}_{n+1}& =& \left({x}_{n}+{\lambda}_{n}\right)/4\\ {y}_{n+1}& =& \left({y}_{n}+{\lambda}_{n}\right)/4\\ {z}_{n+1}& =& \left({z}_{n}+{\lambda}_{n}\right)/4\end{array}$$ |

$${\epsilon}_{n}=\mathrm{max}\phantom{\rule{0.125em}{0ex}}\left(\left|{X}_{n}\right|,\left|{Y}_{n}\right|,\left|{Z}_{n}\right|\right)\sim {\left(\frac{1}{4}\right)}^{n}$$ |

$$\begin{array}{ll}{R}_{D}\left(x,y,z\right)=& 3\sum _{m=0}^{n-1}\frac{{4}^{-m}}{\left({z}_{m}+{\lambda}_{n}\right)\sqrt{{z}_{m}}}\\ & \\ & \\ & +\frac{{4}^{-n}}{\sqrt{{\mu}_{n}^{3}}}\left[1+\frac{3}{7}{S}_{n}^{\left(2\right)}+\frac{1}{3}{S}_{n}^{\left(3\right)}+\frac{3}{22}{\left({S}_{n}^{\left(2\right)}\right)}^{2}+\frac{3}{11}{S}_{n}^{\left(4\right)}+\frac{3}{13}{S}_{n}^{\left(2\right)}{S}_{n}^{\left(3\right)}+\frac{3}{13}{S}_{n}^{\left(5\right)}\right]\end{array}$$ |

The method may fail either because it has been called with arguments outside the domain of definition, or with arguments so extreme that there is an unavoidable danger of setting underflow or overflow.

**Note:**${R}_{D}\left(x,x,x\right)={x}^{-3/2}$, so there exists a region of extreme arguments for which the function value is not representable.

# References

Abramowitz M and Stegun I A (1972)

*Handbook of Mathematical Functions*(3rd Edition) Dover PublicationsCarlson B C (1979) Computing elliptic integrals by duplication

*Numerische Mathematik***33**1–16Carlson B C (1988) A table of elliptic integrals of the third kind

*Math. Comput.***51**267–280# Error Indicators and Warnings

Errors or warnings detected by the method:

- ${\mathbf{ifail}}=1$

- ${\mathbf{ifail}}=2$
- On entry, ${\mathbf{z}}\le 0.0$; the function is undefined.

- ${\mathbf{ifail}}=3$
- On entry, either z is too close to zero or both x and y are too close to zero: there is a danger of setting overflow. See also the Users' Note for your implementation.

- ${\mathbf{ifail}}=4$
- On entry, at least one of x, y and z is too large: there is a danger of setting underflow. On failure the method returns zero. See also the Users' Note for your implementation.

# Accuracy

In principle the method is capable of producing full machine precision. However round-off errors in internal arithmetic will result in slight loss of accuracy. This loss should never be excessive as the algorithm does not involve any significant amplification of round-off error. It is reasonable to assume that the result is accurate to within a small multiple of the machine precision.

# Parallelism and Performance

None.

# Further Comments

You should consult (S not in this release) which shows the relationship of this function to the classical definitions of the elliptic integrals.

# Example

This example simply generates a small set of nonextreme arguments which are used with the method to produce the table of low accuracy results.