424 Major L. Darwin. On the Method of 



144' on each side of the axis, has been adopted afc Kew in order that 

 any error in the measurements on the ground glass may produce a 

 small proportional error in the results. But it should be observed 

 that, the smaller this angle, the less will be the errors just discussed, 

 and by lessening the angular movement these errors can be reduced 

 to any extent, but only with a proportional loss in the general 

 accuracy of the results obtained. 



This is not the place to enter into a general discussion on 

 focometry, but a few words to justify the choice of the Kew method 

 may perhaps be permitted. Many of the known means of finding 

 the principal focal length depend in principle on measuring the 

 relative size of the object and the image, and the foregoing remarks 

 on the errors involved are more or less applicable to them, thus 

 showing that they are open to the same criticisms on theoretical 

 grounds as the work at Kew. Many methods of focometry have to 

 be rejected because they do not measure the distance from the nodal 

 point, and others are unsuitable because the calculations or successive 

 adjustments involved, render the operation too lengthy. There are 

 no doubt many instruments, as for instance that devised by Professor 

 Silvanus Thompson, which do give the true focal length as measured 

 on the axis with theoretical accuracy, but these have not, as a rule, 

 been specially designed for photographic lenses. One method, 

 which is hardly open to criticism on theoretical grounds, may be 

 mentioned in a little greater detail as being that specially recom- 

 mended by the International Congress of Paris ; this is the elegant 

 plan which Commandant Moessard proposes to carry out by means of 

 his instrument, called the Tourniquet, which is described in Wallon's 

 'Traite elementaire de 1'Objectif Photographique,' and elsewhere. 

 Advantage is taken of the principle that if a lens is revolved about 

 an axis passing through the nodal point of emergence, the image of a 

 distant point will not appear to move if seen through a fixed eye- 

 piece ; thus, by successive adjustments and trials, the lens can be so 

 placed that an axis does pass through the nodal point; and, by 

 measuring the distance between this axis and the focus of the eye- 

 piece, the true focal length can therefore be obtained. Since 

 a movement can be detected before it can be measured, a smaller 

 angular movement is required with this method than with the Kew 

 testing camera, and therefore, as far as distortion is concerned, 

 greater, but not absolute, theoretical accuracy is obtained. As for 

 the coincidence of the visual and actinic centres of similitude, better 

 theoretical results are only obtained by this method on the assump- 

 tion, which is probably a true one, that these points approach each 

 other as the point of observation gets nearer the axis. 



By taking observations some 14 away from the axis of the 

 lens, we conclude, therefore, that we obtain the most rapid and 



