236 M. E. Pringsheiin on a Measurement of Wave-lengths 



the ultra-red portion of a diffraction-spectrum. As therein 

 the extreme rays of the spectrum of the first order coincided 

 with the light-rays of the spectrum of the second order, he 

 could easity infer from the known wave-length of these light- 

 rays that of the extreme heat-rays. He found for it the 

 number 0*001075 millim. 



Yet we have no right to regard the rays thus defined as 

 really the extreme heat-rays coming from the sun to the earth ; 

 the number thus found has only an individual significance 

 relative to the particular experiment made, and teaches us 

 nothing further about the sun's radiation than that rays having 

 a wave-length of 0*001075 millim. are found therein. Hence 

 the investigation is by no means concluded, and it is desirable 

 to identify rays of still greater wave-length in the solar spec- 

 trum by fresh observations. Such an investigation I carried 

 out in the course of the summer of 1882, in the Physical 

 Institute of the Berlin University. 



I. Apparatus. 



As the thing required is to verify the presence of rays 

 whose intensity appears to be very slight, the success of the 

 investigation depends essentially on two main conditions. 

 Namely, first, care must be taken that the intensity of the 

 spectrum observed is as considerable as possible; arid, secondly, 

 an instrument must be employed for the demonstration of the 

 extreme rays sought which possesses the greatest possible sen- 

 sitiveness for the class of rays in question. 



To produce a spectrum of the utmost intensity, I employed an 

 excellent plane metallic reflecting diffraction-grating, prepared 

 by Chapman with Rutherfurd's machine, possessing a square 

 aperture of 43*3 millim. side, and having, according to the 

 maker's statement, 17,296 lines to the inch. This gives for the 

 distance d between each two lines of the grating 0*00146852 

 millim. 



In order to control this datum, observations of the lines D 

 and F were undertaken with the grating; and from the known 

 wave-lengths of these lines the mean value of d was found 

 = 0*0014849 millim.; so that we can pretty safely put 



d= 0*001485 millim. 



In order to utilize the greatest possible number of rays for 

 the production of the spectrum, all the sun's rays reflected 

 from the metallic mirror of a Duboscq heliostat were concen- 

 trated, by means of a circular silver mirror of 90 millim. dia- 

 meter and about 200 millim. focal distance, upon the square 

 aperture of a slit, of 1*3 millim. side, placed in the focus of 



