CLARK: HARVARD DEEP-SEA THERMOGRAPH, 523 



a frame carrying many thin blackened metal plates, equally spaced 

 and parallel to each other, the arrangement being similar to a 

 window-blind. The plates are § inch long, .003 inch thick, and .04 

 inch apart. The light when it has passed between the plates is 

 nearly parallel, considered vertically, but is still diffuse in its hori- 

 zontal aspect. It then passes in succession through another slot, the 

 thermometer-stem, a third very narrow slot (.008 inch), and falls 

 upon the film. The stem is orientated so that its eccentric capillary 

 lies very close to the narrow slot which throws a slit image of the 

 capillary and its mercury column on the film. A specimen film is 

 shown, Plate 2, fig. 7. The thin white line in the middle of each 

 record is the shadow of the mercury. The double space near the 

 middle of the film is made once an hour by the clockwork, to facili- 

 tate counting. 



Thermometer. 



The thermometers, two of which are shown, Plate 4, fig. 12, were 

 made by Henry J. Green of Brooklyn, N. Y. They are made of spe- 

 cial tubing j^ inch in diameter, having the eccentric flattened capillary 

 required by the special optical system. Since the temperatures aboard 

 ship will often be well above the highest water temperature, an ex- 

 pansion-chamber is provided at x. The bulb is | inch in diameter, 

 and may have any length up to 3| inches. Thus although the optical 

 system demands a broad capillary (.01 inch), it should be possible by 

 a proper choice of tubing to make a thermometer of greater sensitive- 

 ness than will ever be needed. The photographic record includes 

 nearly four inches on the thermometer-stem. If it may be read to .01 

 inch we shall expect an accuracy of .1° over a range of 36° or of .03° 

 over the 10° range which will be used in deep water. The stems are 

 not graduated. The temperatures are found from the photographic 

 record by measuring the distances between the top of the mercury and 

 the upper end of the image. If desired a fiducial mark could be made 

 on the thermometer-stem from which measurements could be taken. 

 No appreciable error is introduced by distortion of the film in develop- 

 ment. The calibration is accomplished by allowing the thermograph 

 to record the temperature of water in which is placed a standard 

 thermometer. As shown in Fig. 12 the thermometer is mounted in 

 a steel sleeve, being held firmly in place by a packing gland containing 

 cotton cord and shellac. A slot cut in the sleeve makes it impossible 

 to orientate the thermometer wrongly when assembling the apparatus. 



