346 Tlie Lambdh Ohservatory. [July, 



for convenience sake, so placed upon the wall that the angular 

 intervals included between them on the circular arc are respectively 

 30, 60, 120, and 150 degrees. This arrangement, by varying the 

 pairs of collimators used, allows twelve different angles to be tested. 

 The test, in plain language, is simply reading these angles off as 

 they are given by the circles of the mstruments, and then shifting 

 round the circles again and again, so that exactly the same work is 

 performed in succession by different portions of the same graduated 

 rim. If the graduations of the circle are correct and trustworthy, 

 the angles read off between any given pair of colhmators will 

 obviously always have the same value whatever portion of the circle 

 be employed in the reading. This test is a very accm-ate and a 

 very severe one. The slightest inexactness and failure in the 

 mechanical work at once becomes glaringly ob^aous, and can be 

 estimated as a question of amount as well as of fact. In practice 

 the process is repeated with each instrument on successive occasions, 

 to test permanence as well as exactness of construction and per- 

 formance ; — to see that there are no weak and yielding points, or 

 shifting screws, or attachments, involved in the structure. If the 

 reference-points were simply so many fixed spots, established upon 

 the circular wall, it would be necessary that every graduated circle 

 placed under examination should be truly centred upon one point 

 with the most refined exactness. This necessity is practically 

 avoided by placing the reference-points in tubes behind, or beyond, 

 curved lenses of glass, which have the useful property of always 

 sending parallel rays to the telescope of the tried instrument, and 

 therefore always securing the invariability of each angular instru- 

 ment however the centre of the reading circle may he. Colonel 

 Strange has satisfied himself by actual trial that the centre of a 

 graduated circle, or arc, may be shifted a fifth of an inch without 

 making any appreciable alteration in the value of the angles read. 

 It is this especial virtue which converts a mere visible fixed point 

 of reference into what is technically known as the " collimator." 



The testing of vertical circles is a far more difiicult piece of 

 business than the trial of horizontal ones, for this reason — in 

 vertical circles the telescope which is employed in making each 

 observation is rigidly connected with some unalterable radius of the 

 circle, and any given angle included between reference-points cannot 

 therefore be apphed over and over again to different portions of the 

 circle. This however is not of very much moment, because the 

 vertical and horizontal circles of any given maker are turned out 

 by the same graduating instrument, and if the horizontal chcle is 

 found to be graduated correctly, the vertical circle may with some 

 confidence be assumed to be of the same excellence. In the case 

 of the larger and more important instruments the examination is 

 rigidly carried out by observing standard Greenwich stars as they 



