190 SCIENCE PROGRESS 



observations is of the greatest importance. Yet it is surprising 

 that although methods have been devised for determining other 

 instrumental errors which would vitiate the accuracy of the 

 results, little progress has been made towards reliable and 

 accurate determination of telescope flexure. The method 

 most commonly employed, that of determining the horizontal 

 flexure by means of collimators, is not very reliable, and in- 

 volves the assumption, in order to obtain the flexure at any 

 inclination of the telescope, that the flexure varies as the 

 sine of the zenith distance. The use of direct and reflected 

 observations of stars introduces other complications, whose 

 interpretation is not even yet fully understood. Methods 

 have been proposed involving the use of special apparatus 

 attached to the telescope, but they were not convenient inas- 

 much as the apparatus had to be removed when star observa- 

 tions were made. 



A method has been recently devised by Ilmari Bons- 

 dorff of the Poulkowa Observatory, Petrograd, which appears 

 to be free from all objections and has the merit of extreme 

 simplicity. The flexure is determined completely at any 

 zenith distance by an apparatus attached only to the exterior 

 of the cube and having no other contact with the instrument, 

 and which does not need to be removed when star observations 

 are made. 



The essential parts of the apparatus consist of two steel 

 tubes solidly fastened to the cube of the telescope, the one 

 above and the other beneath. The tubes are parallel to the 

 telescope axis. To the object end of each tube a horizontal 

 steel cylinder is attached which fits accurately into brass 

 bearings mounted at the ends of a light but rigid aluminium 

 plate, which when in position is just in front of the telescope 

 objective and perpendicular to the telescope axis. The bear- 

 ings are held against the cylinders by four spiral springs. A 

 mirror, whose position is capable of fine adjustment, is mounted 

 in a brass mount fixed to the aluminium plate. Weights are 

 attached to the eye-ends of the tubes to counterbalance the 

 aluminium plate and mirror. 



It is obvious that in whatever position the telescope may 

 be placed, the plane of the mirror will always be parallel to a 

 plane fixed with reference to the cube of the telescope, for the 

 distance between the ends of the two tubes cannot vary and 





