Hello again I notice that the extend log sometimes includes many START statements. I can understand why the algorithm needs to reset itself once it gets to the end of a branch, and then backtracks all the way back to the original start, having generated all of the side passages on that branch on its way. At this point it needs to decide where to start again. This suggests to me that the user should be able to influence the subsequent start stations, so long as they study carefully where the first branch development is finally exhausted. However, I notice that if I specify multiple start statements, only one of them is ever used. What is going on here, and can the user influence subsequent start stations? In addition some of the start stations appear to be splay stations. Since in general splay stations are not considered part of the primary survey leg centreline structure, surely this is some kind of bug? Here is an extract from an extend log. START 3.67@3.Hawkes START -@2.Hawkes RIGHT 2.29@2.Hawkes done Thanks Bruce