science fair project

The question cloud be: What makes a stone shine? Why does a polished stone shine?

What about showing what happens when you tumble stones with different kinds of grit. You know, a) is a stone in just water b) is a stone in 90 grit c) is a stone in sand... that kind of thing. I'm not sure how complicated first grade science projects get but that would be a good comparison. She could make a guess about which grit would wear a stone down the fastest.

how about why some stones polish and some dont do so well (hardness) why some just crumble and some turn into gems

Here's a project title:

How a rock tumbler duplicates the effect of the ocean upon beach rocks

For those that read this in the future:

To design a great science experiment, start by finding something that can be measured.  Then design your experiment around that.
For rock tumblers, one of the possible things that can be measured is the size of rocks before and after tumbling.
Here's the set-up for Tomahawk's rock tumbling science experiment:

Question:  Which types of rock are most resistant to weathering?

Materials to test: We used concrete chunks (since the kids sometimes found them at school) and local rock samples from a landscaping company.  You can use some of your rough rocks that you usually tumble, but be sure to include other types for contrast.  Local samples can provide relevance, such as something from near your home.  Landscaping companies will often give free samples.  Include a range of hardnesses.  Your rocks should be similar in starting size.

Hypothesis: List your rocks in the order you think will be strongest to weakest, maybe include some brief reasons why.

Method: Measure each rock's longest length.  Tumble the rocks (an equal number of each type, totalling to an appropriate volume for your rock tumbler.)  Take them out each day and measure them again. Then re-set-up your tumbler with fresh grit and continue tumbling.
For older students where scientific rigor matters, you should repeat the experiment three times (tumble and measure a batch of rocks for a week or two, then start over with a new batch of rocks (the same types) and tumble and measure, then repeat the experiment a third time with a new replicant of rocks.)  Younger students can just do the experiment once.

Data presentation:  Calculate average cumulative length lost for each rock type for each day. (Average length of that rock type before the experiment minus average length of that rock type that day.)  Graph time (days) on the x-axis and length lost on the y-axis, with zero at the top, and the lengths going down (since they are losing length.)  Make one color-coded line for each rock type.   I'm including a video link; pause the video to look at the graph to see what I'm talking about.  Graphing length lost instead of length allows all the rocks to start at the same point on the graph: zero.  If they started at different lengths, it would be harder to compare results.

Results:  List the rocks from strongest (least length lost) to weakest (most length lost.)

Conclusion: Discuss your results and why they make sense.  What surprised you?  What did you learn? (Tomahawk learned not to buy plastic rock tumblers!)  For any results that did not match your hypothesis, be sure to explain why the rocks were more or less resistant to weathering than you expected.  Do some research to find out!  It works better if your hypothesis was wrong, because then you have more to explain in the results section, and can show that you learned.

Here is a link to Tomahawk presenting his version of this science experiment:

YouTube link - Rock Tumbling Science Fair Project
Check it out!

Also could weigh the various rocks through each stage.

Yes!!
Weight is an even more accurate measurement of how much rock has been lost than length is.
The reason we didn't use rock weight in Tomahawk's experiment, is because one of his rock types was scoria (like pumice, but denser), which is porous, so it holds water in it's holes.  It would be holding a different amount of water each time, which would mess up the results (we'd be weighing rock plus water instead of just rock.)  The way to accurately measure the rock weights would be to oven-dry them before each weighing.  But there's no time to be oven-drying AND tumbling all day.  So we went with length to save time.  (Also good if you already own a ruler, but not a fine enough scale.)
If you have access to a scale (postal scales are available inexpensively) and your rocks are not porous, use rock weight!  Just dry them off with a towel each time before weighing.  No need to take averages.  Just weigh all rocks of a type together on the scale at once.  (So, all your granite together each day, all your sandstone together each day, etc.)