The regulator should be carefully fitted around the dome and all attachments in connection should be tightly fitted to the plate or bridge so that they will remain rigid when regulation takes place.
The tension around the dome should be even and if a tension spring is used in connection it should be strong enough to keep the regulator against the screw constantly without sticking at any point as the screw is moved forward and backward.
It should also be closely examined to see that there is no shake. This can be determined by lightly taking hold of the segment holding the regulator pins and moving it up and down and side ways before the tension spring is fitted. This should be examined with a glass and a correction made if any looseness is noted.
77. _Effect of the Middle Temperature Error._
In the final regulation of watches it is important that the middle temperature error receive due consideration. This error is always a few seconds fast as explained in temperature adjusting Chapter V, No. 21, and is of some consequence in the larger number of complaints regarding losing rates in the pocket, compared to complaints of gaining rates.
The position rating as well as the final regulation is generally done in normal temperature which produces a rate from two to four seconds faster than the heat extreme and it is to be expected that the pocket rate will be slower because the temperature will be higher than normal. This loss may not be the full amount of the middle error as it would depend upon the actual temperature encountered for the entire twenty-four hours and the watch may only be subjected to the pocket temperature for a part of this period. This works in exactly the same way in a lower temperature, as the variation is a loss in either direction from the middle or normal temperature and in case that the watch should be subjected to a freezing temperature at night the result will be a loss during that period.
As an example we will assume the regulation of a watch in which the temperature rate at the extremes of 40° and 90° Fahr. is perfect, while at the temperature of 70° it will time four seconds fast.
Now if this watch is regulated to no variation in the normal temperature it will be plainly seen that there will be a loss of four seconds per day if the watch is placed in service at either of the temperature extremes. If it had been regulated to run four seconds fast in the middle or normal temperature it would time more nearly correct in the pocket.