32 OEEPE, COMMISSION A, SOLAINI AND TROMBETTI
| | für Eich- und Vermessungswesen of Wien: each of them would have carried out the
adjustment of all the strips, with the Verdin’s and the Van der Weele methods respec-
tively. Then Prof. Zeller took care of the strips adjustment with a method derived from
the Zarzycki’s one which is known as the Eidgenossische Technische Hochschule (E.T.H.)
method. The Commission asked both the Centers to repeat the calculations with the Ver-
din’s method in order to find the resulting calculation errors and to adjust the strips
whose computation had been previously carried out with different methods. However the
| contribution of both the Centers to this plan of experimental work has been rather
| | limited. Therefore we lacked the material needed for a comparative examination of the
different adjustment proceedings.
The comparison is confined to only three methods: two of them ary very similar and
| I have generally given identical results. It is impossible to examine here in detail all the
| | results: we shall, therefore, outline only a few summarizing considerations.
hi As said before, the results of the adjustments made with the Verdin's and Van der
| Weele methods, are practically identical in the triangulations where bz = 0; usually the
WM residual errors are rather small because they are kept within 20 m. It is actually almost
impossible to expect more than that from triangulations 100 km long, without auxiliary
data. At any rate there are some cases where the residual errors, especially in height,
"il reach even 40 m.
In the statoscopic triangulations the Verdin's method gives results which can be
compared to those obtained through a triangulation where bz — 0 as far as the X and Y
co-ordinates are concerned: this is quite obvious since the statoscope does not influence
| | the planimetry. On the other hand, the residual height errors after the adjustment are
| | large and in several cases larger than the gross errors after the linear adjustment. In
| | fact, Verdin's method gives adjusted height values which are independent for the sta-
| | toscopic values introduced into the plotting instrument. To date Van der Weele’s method
| | has not been applied to any statoscopic triangulation by the Commission.
Anyway, the results obtained show that the use of 3rd degree formulas for the alti-
metrical adjustment, leads to very large residual errors; this is a consequence of the fact
that the error propagation laws are completely different from the free triangulation
laws. Analogous considerations can be made on the solar method.
Unfortunately, it is difficult to make a comparison between the ETH, the Verdin’s
and Van der Weele’s methods, since the first one makes use of intermediate points and,
therefore, its results are usually better than the ones which are obtained when we use
some points at the beginning or at the end of the strip. In the few cases, where we have
carried out the adjustment with 3rd degree formulas and intermediate points we have
obtained results which can be definitely compared to those of the ETH method.
The general conclusion we can draw from the comparison we made of the adjustment
hu methods is that we can seldom improve the triangulation results of simple strips by the
use of very elaborated adjustment methods since the adjustment can reduce the errors
only in a limited way. The study of a more effective adjustment method does not seem
possible and the progress of aerial triangulations must be found in the improvement of
the observation methods.
5. A study was also been carried out on the variation in q of each plate of two suc-
cessive pairs (as much for grid triangulations as for real plate): it has the purpose to
| show systematic and accidental errors in one element of outer orientation of the photo-
| | grams.
| The available statistical material was not great, but the one obtained by the brid-
| gings of grids has demonstrated the existence of a real systematic error due to bridge
| base behaviour (its deformation in connection with base sign, its displacement not pa-
| rallel in the space).
