Gaussian Numerical Integration of a Function Depending on a Parameter

All three formulae use twenty-seven points per subinterval. With h = \, (5.3) gives the approximation and lower-sign formulae give errors of —0-000 011 343 and 0-000 007 835, respectively. For h = \, when each formula uses 216 points, (5.3) gives an error of 0-000 000 067 and the two formulae (4.11) give errors of 0-725 524 912 with error 0-000 034 104, while the upper 0- 002 and 0-000 000 003, respectively.