19 



owing to the difference existing in the expansion and contraction of glass and brass 

 at different temperatures, a spirit level so mounted may sometimes become loose, in- 

 volving inaccuracy and unreliability of adjustment. Upon finding that the adjust- 

 ment of a spirit level in an even temperature is not as stable as desirable, the level 

 fastenings, tube, screws, etc. should be examined, to see if any of them are loose. If 

 the trouble is in the screws, tighten them up; but if the spirit level can be shifted in 

 its tube by a touch of the finger, take it apart ; soften the plaster of paris in water, 

 and remove it with a sharp pointed stick of wood. Cautiously move the spirit level 

 with your finger, at first only a trifle to and fro, increasing the length of stroke little 

 by little, until it can be safely taken out without breaking; clean thoroughly. Cut 

 pieces of white paper, of the width of the radius of the tube, and somewhat shorter 

 than the length of the spirit level, but longer than the opening in the brass tube, and 

 insert these of sufficient quantity at the bottom of the brass tube, to fill up the space 

 intervening between the glass and the brass tube. The uppermost layer of paper 

 should, however, be so wide, as to envelope the spirit level up to the opening in the 

 brass tube. Now insert the spirit level, taking care not to touch the glass ends that 

 are sealed up, and place the division or other marks, indicating where the level has 

 been ground to a true curvature, uppermost in the brass tube. The level must be 

 pushed in with sufficient friction to prevent slipping in the tube, yet not so tight as 

 to cause a crack at a subsequent low temperature, as brass will contract more than 

 glass. No part of the spirit level should touch any part of the metal tube. Now 

 prepare some plaster of paris with water, of the consistency of pasteT and pour in at 

 each end enough to fill up the space between the end-pieces and the glass, stirring 

 it sufficiently to make a perfect contact by it and the glass and the brass, but leaving 

 the spirit level ends exposed. Now put the level together, and adjust as described 

 elsewhere. 



There are other causes, such as centers and flanges that have been bent by falls, 

 etc., or that have been worn out unequal expansion or contraction in different tem- 

 peratures of the metals employed in the construction of an instrument, or a non- 

 symmetrical lenghtening or shortening of the air-bubble at different temperatures all 

 of which, singly or combined, tend to impair the adjustment of spirit levels on 

 instruments. Of these we will not speak here, as it requires a most thorough mecha- 

 nician and instrument-maker to trace the cause to its proper source. 



Being assured that the level is mounted as explained above, our advice is, not to 

 meddle too frequently with the adjustment of a spirit level. Though it may appear 

 to be out one day, it may be in perfect adjustment other days. It is the function of 

 a spirit level to indicate the changes taking place in an instrument, so that the 

 engineer may make proper allowance and apply his corrections, as the character of 

 his work may require. The finer an instrument, the more sensitive the spirit levels 

 must be, in order to admit of corrections to arrive at closer results. As a rule, a 

 spirit level that does not indicate changes taking place in an instrument, is too in- 

 sensitive for the character of the instrument, and in many cases entirely unfit for 

 reasonably good work. 



Replacing Broken Cross-Wires. 



The cross-lines in our telescopes are bonafide spider webs (except where plati- 

 num wires have been specially ordered). In case they should be broken, they may 

 be restored in the following manner : clean the reticule frame of all foreign matter ; 

 put it on a sheet of white paper with the cuts on its surface uppermost. Prepare a 

 little shellac by dissolving it in the best alcohol and waiting until it is of the con- 

 sistency of oil. From the spider's cocoon, (those from a small black wood-spider 

 preferred), which the engineer has prudently secured at some previous time, select 

 two or three webs, each about two inches long and of the same appearance. Attach 

 each end of these webs to a bit of paper or wood to act as weights, and immerse 

 them in water for five or ten minutes. Kemove one web from the water, and very 

 gently pass it between the fore-finger and thumb nails,- holding it vertically to re- 

 move any particles of moisture or dirt. Stretch the web carefully over two of the 

 opposite cuts in the reticule frame. Fasten one end by a drop of the shellac, let 

 fall gently from a bit of pointed wood or the blade of a penknife. Wait a moment 

 for this drop of shellac to harden. See that the web is stretched tight across the 



