Hi Anne
I am just wondering what your overall objective is and whether you want to be creating polygons for samples or locations. I doubt you know what I mean as locations are not well documented.
You probably have got the idea that a sample is a record of who went where and when. The 'where' bit of that is typically a grid reference obtained by clicking on a map. However, if you are making repeat samples at the same place, you can define a location record, which could be a grid square or a more complex polygon, and just have the sample refer to that.
So, if you were doing repeat bird counts in local parks you might define the park as a location, each visit would be a sample referring to the park location and all the species counted would be occurrence records with the count stored as an occurrence attribute.
You would use a location for a boundary that you want to define once and use many times.
On the other hand, I can imagine if you were doing a habitat survey instead, you might want a sample to say on a particular date an area was of a certain type. In that case you would want the polygon to be part of the sample as it would be different on a subsequent visit.
Is it this latter case that describes what you want to achieve?
Jim Bacon.