I opened a connection to the streaming API the night of the Canadian election, using a number of filter terms. Most of the time, I was not limited, but I was limited at a few peak moments, getting the standard limit message. This was presumably because my filtered stream exceeded the one per cent of the firehose threshold at those moments. I was looking at one minute of data afterwards, and noted the timestamp_ms field existed in both the tweets I received and the limit messages. If the timestamp for a limit message is between the timestamps for two tweets, does that mean that the missed tweets would also have had timestamps between the two tweets I got, or is that simply a timestamp indicating when the limit message was generated?