de certains essais sur la précision possible à obtenir et propose une forme, nommée la triangulation 
radiale mécanique de précision (TRMP). 
Les résultats obtenus montrent que la TRMP, correctement appliquée, peut donner une précision 
horizontale suffisante pour la majorité des levers sauf ceux à grande échelle pour les plans 
cadastrais etc. 
La TRM ne donne pas de z-coordonnées, mais d’ordinaire l’exigence de précision dans la carte 
définitive demande des nivellements terrestres des points de contrôle pour presque tous les modèles 
stéréoscopiques. 
L’auteur a construit un appareil pour la TRMP et un pour la TRM simple. Les appareils 
seront montrés en connexion avec le VIII: ième Congrès de Photogrammétrie. 
M. van der Weele (Pays Bas) demande si M. Fagerholm pourrait donner une 
indication de la précision relative qu’on peut attendre de l’application de sa 
méthode. M. Fagerholm répond que cette précision est de l’ordre de 0,1 mm. 
Answering a question of Mr. van der Weele, the author says that the accuracy of the relative 
position of points is of the order of 0.1 mm. 
Ensuite M. R. Roelofs (Pays Bas) présente sa communication: »The future 
of radial triangulation». Résumé: 
In this paper the possibilities of numerical radial triangulation, disregarded 
for a long time, are reconsidered in the light of new theoretical and practical 
investigations and modern instrumental developments. 
When dealing with the theory of radial triangulation the first thing to do is to 
investigate the size and influence of systematic errors, which are due to both 
ground height differences and camera tilt being different from zero. 
Several authors have already determined the systematic error in a single 
measured direction or angle, but what is much more important is the systematic 
error in those combinations of directions or angles which propagate through the 
strip, namely; the azimuth transfer and the scale transfer. 
It has proved possible to deduce relatively simple formulas for these system 
atic errors in principal point triangulation, expressing them in the ground 
inclinations of the radiais concerned and the lateral and longitudinal tilts of the 
camera. 
These formulas were applied, as an example, to a number of fictitious flights 
over various types of terrain, the ground inclinations being read from topo 
graphic maps and the camera inclinations being borrowed from the aerotriangula- 
tion of a real strip. The results show that with principal point triangulation the 
systematic errors in azimuth and scale transfer are not greater than the acci- 
cental errors in spatial aerial triangulation for all types of terrain from flat to 
rather hilly country, provided that the average camera tilt is kept well below I o . 
The formulas mentioned need only a slight modification to represent the 
standard accidental errors in azimuth transfer and scale transfer in nadirpoint 
triangulation inasfar as generated by the errors in determining camera tilt. 
Applying them to the same strips as before results in the conclusion that for 
flat and hilly country these standard errors are considerably less than those in 
spatial triangulation and that they come up to them only for mountainous 
country. In these computations it was allowed for an accuracy (standard error) 
of 10' in the determination of camera tilt. 
Concluding it might be said that large areas on earth can quite satisfactorily be 
triangulated by the radial method. With a view to that, Wild, Switzerland brings 
a new type of radial triangulator which is most original in that the pictures, 
which are observed stereoscopically, are fixed in position, while a rotating row 
of measuring marks is the means for observing the angles between radiais. 
Die Zukunft der Radialtriangulation 
In diesem Artikel werden die Möglichkeiten der Radialtriangulation von neuem im Lichte 
neuer Untersuchungen und moderner instrumentaler Entwicklungen betrachtet. 
Insbesondere werden die systematischen Fehler in der Azimutübertragung und der Masstab- 
übertragung bei Hauptpunkttriangulation betrachtet. Konkludiert wird, dass diese Fehler nicht 
Discussion. 
R. Roelofs: The future 
of radial triangulation. 
(Pubi. Ili R 2)