The development of normal distribution

I have had a dream of this from the birth of my grandson ( he is now 6  and half  years old). How to show him concretely  the development of normal distribution.
  I have tought statistics in the university over 30 years and I have had thousands of students.  Now when I am retired I have only one student left, my grandson. 
   I am now alone in the family house far from Helsinki a few days.  I wondered where to find a glass for this treatment.  The ideal glass must be wide  but not too high. I found a really special glass and ideal for this purpose: the windscreen of my grandfather's T-Ford from 1927.  (You can see on the top the manual windscreen wiper, which is surprising short).
   Now when there is none to disturb my experiments, I found from the food cupboard a bag of macaroons.  I counted 250 of them and put them to a cup.  Then I dropped the macaroons between the class and a piece of plywood and I got a decent distribution.  Naturally when using real data and only one experiment, the distribution  may not be perfectly symmetric. But that happens with real data Of course the way how you drop the macaroons may be the reason  that the distribution was not quite  symmetric.
  When my grandson comes here in the summer I will  show this to him.  I will paint one macaroon with special colour for my grandson and another macaroon with different colour for me.  Then if we measure where our macaroons stay compared the mean of the distribution, the distribution of our macaroons will also  follow the normal distribution.
To do that we could mark the points to a paper. The points might be for instance different animals. When we would repeat our experiment  we would find  that thee would be only very few huge animals and very few  very small animals and most would be of average size.
     Many phenomenons as such obey normal distribution, for instance  the distribution of length or weight in the population and