104 
ASTRONOMY: HALE AND ELLERMAN 
their upper extremities appearing at the surface as granules; and that 
in the spots we only see the general structure of the photosphere, as if 
in section, owing to the filaments being here inclined." 
Referring later to the extremely fine filaments, estimated at not over 
0'^03 in diameter, which are sometimes seen on the umbra, Langley 
remarks on their resemblance to the filamentous structure depicted in 
spectroscopic drawings of the chromosphere. To this resemblance, which 
is shown by recent observations to extend to many additional phe- 
nomena, we may now direct our attention. 
It has long been known that the surface of the chromosphere, as seen 
at the sun's limb with a spectroscope, commonly appears as a series of 
slender filaments like blades of grass, supposed by Secchi to correspond 
with the grains of the photosphere. To study the structure of these 
filaments in projection against the disk Ave may utilize the spectro- 
hehograph, which permits their cross-sections to be photographed at 
several different levels. Thus, by setting the camera slit on the Hi 
or Ki line, which represents the low-lying calcium vapor, we can deter- 
mine the size and form of the cross-section at a level below that seen 
visually in the chromosphere at the limb. The calcium lines H2 and 
K2 represent a somewhat higher level, illustrated in figure 3, Plate I. 
At this elevation the minute structure is similar to that of the photo- 
sphere, but the average size of the small bright fiocculi is greater, if we 
may judge from a comparison with the grains in Langley' s drawing, 
reproduced in figure 2 on the same scale. The smallest calcium fioccuH 
photographed, however, are less than V in diameter, and thus do 
not differ greatly in size from the average photospheric grains.^ 
Spectroheliograms made with light from the center of the Ha line of 
hydrogen depict a still higher level, shown on the same scale in figure 4, 
Plate I, and on a smaller scale in the stereoscopic picture (fig. 5, Plate 
I). Figure 4 is enlarged to a scale of nearly a meter to the sun's diame- 
ter from an excellent photograph made under almost perfect seeing 
with the new 13-foot spectroheliograph, recently built in our instrument 
shop for the 60-foot tower telescope on Mount Wilson. The spectrum 
is that of the first order of a grating by Anderson, used with a plane 
mirror at such an angle as to give a dispersion of 3.6 mm. to the angstrom 
at Ha. As the camera slit transmits only the central part (about half 
the width) of the Ha line, a high level in the chromosphere is represented. 
The diameter of the solar image at the focal plane of the 60-foot tower 
telescope is 17 cm., so that the scale of figure 4 corresponds to an en- 
largement of 5J diameters. 
For some physiological cause it is difiicult or impossible to convey a 
