After having installed the optical system as already described, by 
placing a luminous point, instead of the cross-wire one has put on ihe 
focal plane a sensitized film actuated by a clock movement, in order to 
study the behaviour of the gyroscope during a certain time, through the 
track (where the abscissa represents the time, and the ordinate the angles) 
recorded by the image of the point on the film. The interruption of the 
track represent 10 seconds’ intervals. 
The trials accomplished were the following: 
1) Time employed by the primary axis to assume the exact orienta- 
tion (see diagr. C) — The arrow shows the moment when the primary 
axis is stabilized. This takes place already 180" before that the motor has 
reached its regimen speed. The oscillations when the axis is stabilized do 
not eccede 30" of arc with respect to the base track. 
2) Influence of the transversal accelerations (see diagr. D) — After 
having placed the whole contrivance on a smooth plane, it has been 
given (once the gyroscope was stabilized) transverse accelerations in the 
sense of the primary and of the secondary axis, of the value of 5 and 
10 meters /sec. 
The oscillations as revealed by the diagram remain of the same order, 
and in any way they are not in correspondance with the accelerations (one 
second after the interruptions); one can assume therefore that their effect 
(which by the way should not exists when the flight is correct) is practi- 
cally zero. 
3) Tilt around the primary axis (see diagr. E) — The device was 
given tilt movements around the primary axis, and back to the normal 
attitude for angles of 6? (sexages.) and for times (to and fro) ranging 
from 10” to 1,5”. Both on the reading of the maximum excursion of the 
point, and on the examination of the track after the return to the normal 
attitude, were noticed errors not bigger than 50” of arc. 
4) Prolonged permanence with an unusual attitude (diagr. F) — Even 
with 60" permanence with an axis tilted of 6° the results are the same 
as with the previous trial. As a conclusion one has noticed that even 
with abnormal flicht conditions the instrument does not surpass, as ano- 
maly of recording, 50” of arc for each component; which corresponds, in 
the limit case of both maximum errors, a difference of recording of ihe 
plumb point of 50”x V5 =T70" of arc. 
When one considers that this represents a first sperimental realiza- 
tion, which is susceptible of further improvements, the results appear no 
doubt as encouraging.