844 
and how far, these deductions are verified experi- 
mentally. This is being done repeatedly with the 
chemical theory of the dynamics and statics of the 
hemoglobin reactions. If only those who believe in 
the adsorption theory would make some precise 
deductions from their theory, it would be easy to 
test that also. At present it evades any quantitative 
trial. 
Attempts have been made to apply the Phase Rule, 
and to attribute the properties of large-scale matter 
to the single ultimate unit of hemoglobin as it exists 
in solution, Presumably this ultimate unit has a 
diameter about 10 times that of the oxygen molecule : 
it is presumably in violent, oscillatory (thermal) 
movement; there is no good evidence that it has 
ever been observed with the ultramicroscope. To 
regard it therefore as a separate phase is to disregard 
the statistical basis of the Second Law from which 
the Phase Rule is deduced. Jf the hemoglobin unit 
be indeed a separate phase, then admittedly the 
known number of degrees of freedom of the hemo- 
globin-oxygen system prohibits the possibility of 
regarding oxy- and reduced hemoglobin as separate 
chemical compounds. No evidence, however, can be 
given for the existence of hemoglobin, in solution in. 
water, as a phase separate from the water, except 
that it can be precipitated by various violent means— 
which surely is not evidence ; the separate phase is a 
pure hypothesis and must be judged by its fruits, 
which at present are difficult to discern. 
Sir William Bayliss’s attitude of continual and 
friendly scepticism, on this particular subject, has 
had one important and valuable effect, the effect 
which he set out to achieve, which, however, his 
modesty prevents him acknowledging, or possibly 
even from appreciating. It has urged a number of 
workers to produce, what was badly needed, a body 
of sound quantitative experimental evidence on one 
of the most fascinating problems in the borderland 
between biology and chemistry. The evidence is 
not complete and we cannot convince him yet; but 
if he will only maintain his scepticism, in an equally 
friendly way, for a few years more, he will really 
force us to produce all the testimony which he 
requires. A.V, Him 
The University, Manchester, 
May 31. 
In the recent correspondence touching the nature 
of the combination of hemoglobin with oxygen, 
references have been made to Wo. Ostwald’s adsorption 
theory. It may clarify the issue if I remind readers 
of NaturRE what that theory was. Wo. Ostwald 
argued that the equilibrium between oxygen and 
hemoglobin could be expressed by a curve based on the 
following equation, X =KC™, where X is the amount 
of oxygen combined with the hemoglobin, C the con- 
centration of oxygen in solution, K a quantity pro- 
portional to the total mass of haemoglobin present, 
and ma constant. The graphic expression of this 
equation must necessarily be a simple curve which is 
at all points concave to the abscissa. No published 
curve representing the equilibrium between hemo- 
globin and oxygen, which has been determined 
experimentally, is of this character, all being more 
or less S-shaped, though in some cases the convex 
inflection is very slight. 
It may seem strange that a theory should have 
been put forward which is at variance with the facts 
in so fundamental a respect. In justice to Wo. 
Ostwald it must be pointed out that he wrote before 
the experimental technique now in use had been 
elaborated. The most recent curves at his disposal 
NO. 2799, VOL. IIT] 
NATURE 

[JUNE 23, 1923 
were those of Bohr, Hasselbalch, and Krogh (for the 
oxygen hemoglobin equilibrium at various CO, 
pressures), These are S-shaped in character, but at 
the time commanded less confidence than they 
deserved ; I think because they were determined not 
as individual curves but as a surface in three dimen- 
sions, the published curves being contours. All 
modern work has confirmed the essential character 
of the curves of Bohr, Hasselbalch, and Krogh. 
Finally, may I pay a tribute to the helpful nature 
of Sir William Bayliss’s criticism (NATURE, May 19, 
p. 666), and suggest an extension of that help in the 
direction of his modifying Ostwald’s theory, expand- 
ing it into an equation which would fit the facts — 
sufficiently exactly to stimulate further research on 
the subject. J. Barcrort. 
Physiological Laboratory, Cambridge, 
June 6. 

In his letter published in Nature of May 10, Sir 
William Bayliss suggests that two cases of adsorption 
do not come within the definition of adsorption to 
which I directed attention in Nature of April 14. 
These are the cases when two or more substances are 
adsorbed upon a surface, and when a substance is 
adsorbed to a thickness of several molecules. Both 
these cases were intended by me to be included, and 
I think reasonably so, with the definition that it is a 
case of adsorption, if the substance is taken u 
uniformly over the whole surface; uniformly, that 
is, when the scale of measurement is large compared 
with individual molecules. This sense of uniformity 
is well understood in the theory of gases, where a 
mixture of gases or a single gas may be said to fill 
space uniformly, with equal correctness. I had no 
intention of limiting the definition to layers only one 
molecule thick ; indeed perhaps I may be permitted, 
as it is suggested that I accept Langmuir’s views, to 
point out that the theory employed by Langmuir 
does not seem to me necessarily to postulate that 
adsorbed layers are always one molecule thick. Such 
a proposition could only be established by estimating 
the amount adsorbed on unit area and calculating 
the thickness of the layer in terms of known data as 
to the size of the molecules in every case of adsorp- 
tion ; it does seem to be established by the beautiful 
experimental work of Langmuir in many cases, but 
is not, I think, claimed by him to be an invariable 
law governing adsorption. 
Sir William Bayliss says in his first paragraph that 
no serious attempt has been made to consider surface 
phenomena in the combination of oxygen and 
hemoglobin, since Wo. Ostwald showed that the 
data of the taking up of oxygen by hemoglobin 
could be expressed by the adsorption formula; but 
he seems to have overlooked that the sole argument 
put forward in my letter of April 14, to prove that 
the attraction of hemoglobin for oxygen is a highly 
localised property of the hemoglobin particle, was 
that the hemoglobin is so much larger than the 
oxygen with which it combines that the oxygen 
must be attached to only a very small portion of the 
surface. If there were general attraction of the 
surface of the hemoglobin particle for oxygen, then 
combination would not stop when only a small 
fraction was covered, but hemoglobin would take 
up much more oxygen than it actually does. Surely 
this is a very definite attempt to consider the surfaces 
of the particles. It is a mistake to confuse the 
argument used in my letter with those based on the 
well-known mass-action formule of Barcroft and 
Hill; it is entirely independent of them, and essenti- 
ally treats the hemoglobin in solution as a hetero- 



