is to avoid the need of thinking in any form during a sequence of operations. The 
operator’s role may be to shift a brush to the next position but should not be to 
think out the next operation. 
3. Improvements in the Design of the Stereocomparator. 
When confronted with the formidable amount of calculation involved in 
analytical photogrammetry, one is apt to forget that the observations themselves 
take considerable time. Also the errors of reading of the various micrometers, and 
any mistakes which may be incurred in recording them, are bound to increase the 
amount of calculation which follows. It is therefore very desirable that an obser- 
vational technique be developed which reduces the time consumed in making the 
observations and eliminates mistakes. 
The time consumed in making the observations may be broken up into three 
parts: 
The first part is the time taken in making the various pointings. There is not 
much hope in cutting down the length of this time. 
The second part is the time consumed in reading the various scales and 
micrometers and in recording these readings. So far it has been found necessary to 
employ a booker to record the readings called out by the observer. The booker calls 
back the reading to the observer. It seems to me however that by this time the 
observer’s attention would have been shifted to another micrometer or scale; if 
this is so, it would be rather doubtful if this verbal communication between the 
observer and the booker serves the purpose of checking. 
The third part of the time associated with the observational procedure is the 
time taken in abstracting the data, and in adding the various corrections. 
The averaging of observations and the addition of corrections can be mecha- 
nized by the use of a sufficient number of the registers described in Section 2.2. 
So let us now focus on the possibility of reducing the time of reading and record- 
ing of observations. 
It has been suggested that time can be saved and reading errors eliminated by 
projecting the scale readings in such a way that the booker may record them 
without the help of the observer. This is a good idea; but it seems to me that it 
would complicate the design of the apparatus, and, in any case, a booker would 
still be needed. Another possibility is the photographic recording, for instance like 
the recording of some modern theodolites. This also is not entirely satisfactory 
because the time taken in processing the film, and in reading and recording the 
readings may even be longer than the present procedure. 
It remains to examine the possibility of having the readings automatically 
recorded, thereby saving the booker’s time and eliminating errors of reading and 
recording, as well as increasing the production of each observer per hour. 
The automatic recording has proved to have a reasonably simple solution, but 
it required sufficient modifications of the known types of stereocomparators to 
warrant the design of a new instrument. The new instrument would have all its 
essential movements controllable by rotating spindles, because it is far more prac- 
ticable to record automatically the number of revolutions of a micrometer head 
than the reading of a linear scale. The displacements can in this way be recorded 
to the nearest thousandth of a millimetre,