Definite integration error in MMA 3.0

*To*: mathgroup at smc.vnet.net*Subject*: [mg8831] Definite integration error in MMA 3.0*From*: luca ciotti <ciotti at boas5.bo.astro.it>*Date*: Mon, 29 Sep 1997 02:39:56 -0400*Organization*: Osservatorio Astronomico di Bologna*Sender*: owner-wri-mathgroup at wolfram.com

Hello, I found a wrong result performing a symbolic definite integral with MMA3.0 running under Linux. A=Integrate[1/Sqrt[1+t+t^4], {t,0,Infinity}] As espected A is expressed in terms of elliptic integrals, but when evaluated N[A,40]=-2.2405....... that is obviously wrong. NIntegrate produces A=1.66159..... It is this a known problem? ciotti luca osservatorio astronomico di bologna via zamboni 33 40126 bologna (italy)