TIIK ABSORPTION ANO EJIISSION OK AIU. 7 



remaining between the line of grease and the border of the oi»eniiig. Tliis is a 

 matter of cousequence ; because any grease exuding into tlie inside gradually 

 spreads over the entire siii'Lace of the window and poweii'ully aljsorbs the most 

 i-efrangible rays. The state of the film of grease has, however, no greater influence 

 upon the dui'atiou of the air-tight closure than has the form of the bearing surfaces. 

 We know by the behavior of the glass cocks of the mercurial air-pump that cones 

 without openings close tightly for years, even though they be in frequent motion, 

 like the above-mentioned cones of the micrometers and main axes. Less trust- 

 worthy are plane annular and discoid surfaces, particularly if they be subject to 

 frequent movement. Still less secure are rectangular surfaces that are subject to 

 motion of translation. Rut the air-tightuess is diminished in ([uite a surprising 

 degree by openings thi-ough the bearing surfaces such as the passage of light or of 

 the gas to be exhausted may demand. Such openings, as I have found in the case 

 of optically plane surfaces, have a tendency to shove the grease together at the 

 openings, and thus gi-adually to denude the bearing surfaces, and in this way to 

 form a canal into the exhausted chamber; and this canal, though fine at first, soon 

 inci'eases in cross-section, so that the air, which at first enters only in excessively 

 small amount, soon begins to pour in in larger volume, especially if the closing 

 parts continue to be moved from time to time. AVhen this sort of leakage once sets 

 in, there is no other remedy than that of replacing the spoiled greasing by a fresh 

 one. The circumstances most adverse to tightness are where there are plane slides 

 moving rectilinearly over rectangular openings. Here the denudation of the sur- 

 faces from grease at times takes place after but little sliding back and forth ; so 

 much so, that, with some kinds of grease, I never could succeed in getting an air- 

 tight closure. For such snifaces Cliesebrough's vaseline, w^ithont wax, answers best. 

 Slides of this kind ought to be pi-ovided with the largest possible bearing surfaces, 

 such surfaces presenting a longer resistance to the tendency of the air to pour 

 in. Here may be mentioned another inconvenience which regularly occurs as soon 

 as the greased parts come to be frequently moved, as for example is the case with 

 all the cones. The initial facile mobility of these parts is of short duration. After 

 some use they begin to run considerably harder, and whei'e the bearing surfaces 

 ai'e subjected to stronger air pressure they begin to o[)pose no little I'esistance to mo- 

 tion. This inconvenience is most felt in the cold part of the year when the grease 

 is harder. This can only be remedied by frequent greasing of the surfaces in ques- 

 tion. The considei-able resistance which the great cones of the camera and of the 

 prism-block oflEer to turning calls for an installation which shall prevent lateral 

 displacement. Otherwise, the glass tubes leading to the air-pump will be exposed 

 to constant danger of breakage. I have met this need in my spectrograph by 

 bridging over [each of ?] its three levelling-screws with _J~L_ shackles of the 

 form°here shown, so that the screws perfectly permit a vertical, but n<. lateral, 

 motion of the apparatus. A still more detailed description of this spectrograph is 

 contained in ray report to the Imperial Academy of Vienna : Sitzunrjsberichte der 

 hiim-Uclien Akademie der Wimev><dMften in Wien ; mathem.-natunv. Classe, vol. 

 cii, Pai-t Ha, June, 1893, pp. 625-694. 



