When we began our project, our intent was to take multi-flash photographs of several balls and compare the collision durations and compression ratios of each ball.  However, due to time constraints, we were only able to gather significant data for one ball.  This may be a limitation of our experimental methods; if we had to do this again, we would probably look for a way to take multiple photos of the same collision without repeating the collision so many times.  The first ball that we tried was too rigid; its collision was too brief for us to measure with the apparatus we had in place, and we didn't have sufficient time to devise a more appropriate way of photographing it.  When we prepared to analyze the data we collected from the photographs of the green "Goooz" ball - which is the one we collected a great deal of data on - we ran into a problem.  Because the photos were taken in multiple sessions and because the ball was dropped by hand, it was not always exactly the same distance from the camera.  This made it impossible to compare the size of the ball between two different photographs.  By way of solution, we decided to analyze the ratio of the ball's width to its height.  This quantity does not depend upon how far the ball is from the camera; for example, an uncompressed ball will always be as wide as it is high.  We found that, for the green "Goooz" ball, the compression ratio decreased linearly until it reached a minimum 14 milliseconds after impact and then increased in a largely linear fashion until the ball left contact with the surface 32 milliseconds after impact.  At its most compressed, the ball was over three times as wide as it was high.  Our graph of compression ratio versus time could have been either linear or parabolic.  From a physics angle, it's probably more likely that it was parabolic; a linear relationship would mean that the ball was behaving like a spring, which would be unexpected since the ball is very "squishy."