228 HAEMOGLOBIN. 
remain very constant, the increase in the amount of oxyhemoglobin 
influencing more the intensity of the band than its width. In the case 
of solutions containing approximately the proportion of oxyhemoglobin 
above mentioned (i.e. from - 20 to 0"33 parts in 1000), the spectral 
region between F and G is absolutely unshaded. Soret's band is then 
seen, extending from X 404-/. 434; i.e. it occupies the greater part of the 
spectral region intervening between G- and H : the edges, however, 
uniformly shade away as far as these lines. 
By examining a series of photographs of spectra obtained by inter- 
posing solutions of oxyhemoglobin of very different concentrations, I 
have determined that the mean ray absorbed does not, as Soret thought, 
coincide with h (x 410T), but is decidedly on the red side of that line, 
corresponding to X 414. 
When the concentration of the solution of oxyhemoglobin increases, 
the width of the band very slowly increases. Its less refrangible border 
never passes beyond G ; as the solution becomes highly concentrated, 
the band widens perceptibly, and it does so in the direction of the ultra- 
violet. With a solution made by diluting 1 volume of blood to the volume 
of 250 (water or 01 per cent, solution of Xa(OH) being employed as the 
diluent), the absorption-band, though much more intense than with the 
more dilute solutions, retains almost the same boundaries, its shadowy 
borders approaching, but not passing beyond, G and H. "With a 
solution containing 1 part of blood in 100, the appearances differ 
remarkably from those previously referred to. The solution is 
transparent for light from F to nearly G ; it transmits light with 
difficulty from L to N (X 381*9- x350'01) ; the remainder of the ultra- 
violet is completely absorbed. A solution containing 5 per cent, of 
defibrinated blood (or about 6 - 5 parts of oxyhemoglobin in 1000 parts) 
absorbs the whole of the violet and ultra-violet regions of the spectrum, 
with the exception of a region between F and G, but nearer the former 
(X 460-X 490). 
It remains to be considered with how dilute a solution of oxy- 
hemoglobin a photographic record of Soret's band can be obtained. 
Examining a stratum 10 mm. broad I have obtained definite results, 
when the solution contained somewhat, but not much, less than 1 part 
of oxyhemoglobin in 10,000. 
No colouring matter yet investigated exhibits the intense absorption- 
band between G and H which is characteristic of hemoglobin and its 
compounds. Several substances (carmine, picro-carmine, and the colouring 
matter of alkanet root) exhibit absorption-bands in the visible part of 
the spectrum which bear a superficial resemblance to those of oxy- 
hemoglobin. The spectrum of none of these colouring matters exhibits, 
however, any absorption in the extreme violet or the adjacent ultra- 
violet. 
The researches which I have conducted have shown that the band of 
Soret depends on the iron-containing group existing in the hemoglobin 
molecule, yet not upon its iron. The variations in character and position 
which this band exhibits in the various compounds and derivatives of 
hemoglobin will be referred to under each. 
