288 



SCIENCE. 



[N. S. Vol. XV. No. 373. 



duration. (The paper mil appear in full 

 in the American Journal of Science.), 



A Comparison of Printing and Eecorcling 



Chronographs: C. S. Howe. 



A printing chronograph, to be of any 

 value to astronomers, should be easily set 

 and regulated ; it should run several hours 

 without any attention ; it should have a 

 small probable error; its results should 

 agree, or differ by a constant amount each 

 night, with the results from the recording 

 chronograph. Several months' experience 

 with the Hough printing chronograph at 

 the Case Observatory shows that the in- 

 strument can be easily set and adjusted 

 and that it runs several hours without any 

 attention. The probable error of one record 

 was found to be ± .011 seconds, and the 

 probable error of the mean of nine wires 

 ± .004 seconds. Similar tests with the 

 recording chronograph give respectively 

 ± .006 and ± .002 seconds. These errors 

 are so small that they can be neglected. A 

 direct comparison of the two chronographs 

 showed an average difference between them 

 of .025 seconds, the printing chronograph 

 making the record first. 



Tlie Clock Room at the Case Observatory : 

 C. S. Howe and E. H. Brown. 

 The clock room is a room built of tile 

 within one of the rooms of the basement of 

 the Physical Laboratory. The basement is 

 warmed in the daytime, but not at night. 

 The space outside the clock room is warmed 

 by two gas stoves, the gas being turned on 

 or off by an automatic burner which is 

 regulated by a thermostat made of steel 

 and hard rubber. This controls the tem- 

 perature outside the clock room within one 

 degree centigrade and inside the clock 

 room within about one half degree. The 

 inside temperature is further controlled 

 by another thermostat, which by means 

 of a relay throws in or out one or more 

 electric lamps, the heat of which changes 



the temperature. This arrangement keeps 

 the temperature within one tenth of a de- 

 gree. The hourly rates of a Riefler clock, 

 enclosed in a glass case from which the 

 air is partially exhaiisted, were also given. 



The Almucantar as an Instrument for the 

 Determination of Time: C. S. Howe. 

 In 1885 Dr. Chandler compared the 

 clock corrections determined with a small 

 almucantar with those determined with the 

 large Transit Circle of the Harvard Ob- 

 servatory. The Case Almucantar is a large 

 instrument with a six-inch object glass. It 

 was not possible to compare it with a large 

 transit and so its value for the determina- 

 tion of time could only be determined by 

 the consistency of the several results ob- 

 tained on any one night and by the excel- 

 lent and nearly uniform clock rates deter- 

 mined by it. The differences between the 

 several values of the clock corrections on 

 any one night were usually not more than 

 two or three hundredths of a second and 

 the greatest difference since June, 1901, 

 was 0.13 of a second. 



A Description of the Second {Chile) Mills 

 Spectrograph: "W. W. Campbell. 

 The mechanical features of the instru- 

 ment and its method of support are radi- 

 cally different from those of the conven- 

 tional spectrographs. The entire frame- 

 work is composed of a single steel casting 

 with webs and flanges bracing it thor- 

 oughly in every direction and with especial 

 reference to the supports near its two ends. 

 The slit, the prism box, the camera and 

 the collimator and camera lenses are all 

 fixed directly to the casting. A reflecting 

 slit is used, curved to make the spectrum 

 lines straight. The weight of the entire 

 spectrograph is aboiit 75 pounds. The in- 

 strument is mounted in a supporting cradle 

 composed of tee and channel steel bars in 

 such a way that it will rest on two points 

 (a plane and a ring) near the two ends of 



