Chuck,

You should use smaller resistors for the divider.
This is from the datasheet for a 16F877
The maximum recommended impedance for analog sources is 10 kohm
Your 100k and 50k are much higher and can cause error in the reading due to the sample and hold time requirements.

The example I gave already works with both the "Solar Cell and Lead Acid Battery" so, not much to modify.

And, as far as the DIV32 goes... NO it does not slow things down substantialy. In fact, when you divide 2 WORDs just like you always do in PBP, it actually uses the same DIV32 routine internally to accomplish it. So it doesn't take very much time at all. I've heard that statement before, to the extent of some people saying to "Avoid it at all costs". Rubbish!! Use it as much as possible, and forgive them, for they know not what they say.

Darrel