hi will, bruce made quite a good analysis of your example. i got the same result as he. personally, i found that, while it is convenient to separate scrap and data, it's better to organize the project in small selfcontained chunks (say a few surveys or pieces of them and their maps) that can be compiled (and debugged) individually. next i put the chunks together, creating a survey hierarchy, with bigger and bigger maps, solving the equate and the joins a piece at a time as i go up in the hierarchy. just my way to manage this complexity. there are a few fine points, but this is the big picture. attached is a "revised" archive of your project: the map with the drawings is included (selected) in the output: 1) equate station A2@... with a2@...: not sure this is correct 2) changed the station names in the scrap (just a quick fix) marco