192 



8CIENCB 



[N. S. Vol. XXIX. No. 735 



faint or diffuse bands in spectra of small dis- 

 persion, whence tlie name. After trial of 

 other comparison objects, I conclude that 

 nothing succeeds as well for the purpose of 

 matching one of these hazy bands in a photo- 

 graph of the spectrum as an equivalent, but 

 not identical line or band in another spectrum 

 of identical intensity as to the general back- 

 ground of continuous spectrum. The com- 

 parison line is chosen both narrower and 

 brighter than the one to be measured 

 in order that, when placed a little out 

 of focus by a displacement of the objec- 

 tive of the microscope with which it is viewed, 

 it may appear both broader and fainter, and 

 may thus resemble the object with which it is 

 to be compared. Beyond this, the observa- 

 tion consists simply in repeated settings of 

 the microscope in its out-of-focus position 

 corresponding to apparent equality of the 

 hazy bands, and in the determination of a 

 curve of brightness by a careful photometric 

 calibration of the scale of the instrument. 



The measurement is safeguarded in every 

 possible way, and especially by duplicate 

 measures, made on each one of the plates, of 

 a line which is certainly solar and not subject 

 to modification by the atmospheres of the 

 planets. For this purpose the hydrogen C 

 line has been made a test object, and also a 

 means by which the measures on little a may 

 be corrected for trifling variations in the focus 

 of the speetograph in the intervals between 

 successive spectrum exposures. 



The measures on great G in the spectra of 

 Mars and the moon, taken at Flagstaff by Mr. 

 Slipher at such times as to give equal alti- 

 tudes for the two bodies at midexposure, and 

 for such durations as to prod'uce equivalent 

 intensity of spectral background under identi- 

 cal photographic development, have a ratio, 

 which never departs much from unity; and 

 the average ratio for all of the plates meas- 

 ured approaches unity much within the limit 

 of the probable error. On the other hand, 

 .similar series of measures on little a show 

 without exception that little a is more in- 

 tense in the spectrum of Mars than in that 

 of the moon, with equal C lines and for equal 

 altitudes. 



The average value of the direct readings of 



the spectral band-comparator makes the in- 

 tensity of a in the spectrum of Mars = 1.224 

 =i= 0.0245. The observations in full may be 

 found in Lowell Ohservatory Bulletin, No. 

 36. 



This result requires further correction by 

 means of a calibration curve for the readings 

 of the spectral band-comparator, before it can 

 be stated in absolute units. To obtain such a 

 curve, I have measured the disappearances of 

 the same spectral line at various settings of 

 the instrument with the modification of 

 Pritchard's wedge-photometer invented by 

 Dr. Charles H. Williams, of Boston, and called 

 by him a "simplex" photometer. It is a 

 well-known fact that the eye is much more 

 sensitive to slight variations of intensity than 

 we might have anticipated. But just as we 

 do not recognize our faculty of discrimina- 

 tion, neither do we recognize how great the 

 actual differences of intensities of illumina- 

 tion may be which correspond to our fallacious 

 impressions of the same. Differences of 

 illumination which we note immediately, but 

 which are not judged to be great, which in 

 fact, at first guess, we estimate as but a few 

 per cent., turn out to be many hundred per 

 cent, when made the subject of exact measure- 

 ment. It is probable that very few would 

 guess in the absence of exact measurement 

 that a first-magnitude star is one hundred 

 times as bright as a sixth magnitude. Bear- 

 ing these facts in mind, it need excite no 

 surprise when I state that the actual ratio of 

 intensity of the a band in Mars, when ex- 

 pressed in absolute units, is much greater 

 than would at first he inferred from the direct 

 readings. I find that the real intensity of the 

 band in the spectrum of Mars in the month of 

 January, when the dew-point at Flagstaff was 

 about 20° F., was four and a half times as 

 great as in the lunar spectrum at the same 

 altitude. 



I have endeavored to find out how far it is 

 possible for this result to have been vitiated 

 by inequalities of photographic development. 

 Some of the spectrograms of Mars have been 

 over-exposed, but by a detail in the mode of 

 procedure to which only incidental allusion 

 has been made, namely, by altering the illumi- 

 nation of one or the other of the spectrograms 



