(183) 
the present state of the art. Eventually, we may arrive at the point where the 
navigational system of military aircraft provides the true vertical direction 
with a very high degree of accuracy. When that will happen can not be pre 
dicted, but it is likely that such navigational systems will be very costly, at 
least for a long time to come. I would, therefore, like to skip this possibility 
and suggest postponement of its consideration until more realistic data are 
available. 
If we consider the vertical problem from military viewpoints we can state 
that a true vertical reference will be acceptable only if it does not put addition 
al limitations on the operation of the aircraft. Necessarily, even under combat 
conditions, photography for mapping and charting purposes will have to be 
produced under so-called straight and level flight conditions, as well as these 
are achievable. But the direction of flight lines, and the time at which they 
are to be flown, must be exclusively identified by operational requirements. 
This almost necessitates that the vertical be produced within the aircraft during 
flight. The problem is not new either and, as you all know, numerous attempts 
have been made to solve it in various ways. The most investigated approach 
is that of the artificial horizon represented by a pendulously erected single 
gyro or multi-gyro platform. Although such equipment is widely in use now, 
either for instrument flying or as control equipment for automatic pilots, its 
indication of the true vertical under flight conditions is of very limited accu 
racy. It has been proven by several independant sources through evaluation of 
photographs taken over well-controlled territory that almost 40% of the time 
the indication furnished by a high precision horizon gyro deflects more than 
one-half degree from the true vertical. The gyro remains within one degree 
almost continuously, but occasional excursions of its reference line in high 
speed aircraft must be expected up to one and one-half degrees. This happens, 
under so-called straight and level flight conditions, to a gyroscope which, under 
static conditions in the laboratory, will maintain its reference line within T 
from the true vertical over a practically unlimited period of time. 
Extensive tests have been performed at Wright Air Development Center 
by the Photographic Reconnaissance Laboratory to determine the sources of 
this discrepancy between static and in-flight performance of horizon gyros. 
The proper function of a horizon gyro is based upon the general assumption 
that all accelerations other than gravity which occur during flight integrate to 
zero over a period of time which is reasonably short with respect to the para 
meters of gyro performance. The mentioned tests show that, even under con 
ditions of so-called straight and level flying, this assumption cannot be made. 
A gyro, freely suspended at its center of gravity, theoretically is a means of 
maintaining a position in space. Due to residual unbalance and friction of the 
suspension system, any such gyro has an arbitrary free drift which, for a 
reasonably high precision instrument, amounts to approximately 15 /sec. In 
order to make such an imperfect gyro assume the direction of the vertical, an 
erection system is provided which, activated by a pendulum, precesses the gyro 
into the desired direction. The pendulum employed is of extremely short 
periodicity and is very nearly critically damped. For all practical purposes, it 
will therefore assume the direction of gravity instantaneously under static con 
ditions. Under such conditions, it will erect the gyro to the direction of gravity 
with high precision at a rate of approximately 1 /sec. Necessarily, the rate of