Hello again, After much head scratching and installs on difefrent servers in the same domain, we have managed to get the send port to send. Using the technique on this web page: http://www.traceofthought.net/PermaLink,guid,273a4374-40fc-4fe4-bbe6-833aa3c84022.aspx we seem to have fixed the issue. The implication therefore is that the problem was with the way the WSE Adapter is handling threads and using the thread pool – the WSE Adapter didn’t have a ’free thread’ to process the response from the service.I had read the MSMQ article (linked above) a few days ago and had discounted it from our inquiries on the basis that the processes in question were the only thing running on the host in question (where there is a single WSE Send Port), there is only one receive location involved and we need only submit a single instruction to replicate the issue (in test we successfully tested the Web Service with 100 at once). This had led to a little confusion on my part – given the fact that there was unlikely to be any other threads using the WSE Adapter why did this happen, and why did it only affect Psuedo (we do not need these entries in Dev/Test)? Does anyone know if the fact that the Pseudo box is the first physical server (Dev and Test being virtual) is the best suspect for the difference in how the threads behaved, causing the need for these changes? CheersStu.