A question regarding Loess smoothing and understanding the information concerning the errors on the method

["Gladwin, Philip [FI]" <philip.gladwin@citigroup.com>]

Good morning,

I am using the Loess algorithm in Splus to perform some smoothing.

According to the Insightful S-Plus 6 'Guide to Statistics Vol. 1' page
435 I can obtain some diagnosticse, e.g. Residual Standard Error, and
Multiple R-Squared error.

Could somebody pleases tell me exactly what is meant by
Residual Standard Error
Multiple R-Squared
and residuals 
presented in this example?

I am surprised that these parameters are not described in more detail.

Alternatively could somebody suggest a good place to go to find out?

I am using Splus 2000 Professional Rlease 2 on a Windows 2000 machine.  I dip 
in and out of Splus
every so often and consider myself to be a novice.


kind regards,
Philip Gladwin
FX Analytics