>This is unfortunate, but expected since VARTYPE treats the expression, well as an expression that may fail. While non-existing single (non . vars) will generate "U" results, VARTYPE() will cause an error.
>
>IOW, you can use VARTYPE() in this situation only if you know at
>the very least that all objects actually exist and will return
>some sort of type or NULL.

Thanks for the explanations, gentlemen. We'll just have to be a little careful when using VARTYPE() when object existence is questionable, is the bottom line then (but you should really be checking for existence anyway then, of course). Otherwise it's a fine new addition to the vfp repertoire. There has been some confusion, as I mentioned earlier in thread, since a FP Advisor 6/98 article noted that VARTYPE() would return an error if ANY variable didn't exist...