COMPOSITION OF THE PRO TEWS. 27 
result will be achieved, when proteids obtainable in a crystalline form 
have been thoroughly investigated. 
Vegetable proteids have been prepared in a crystalline form 1 in 
combination with magnesia : Dreehsel 2 found in one preparation 14 per 
cent, of magnesia (MgO): in another, prepared by an improved method, 
L'43 per cent. From this the molecular weight x may be calculated as 
follows : — 
x 100-1-43 ..,_-_ 
— = — ; x = 2,1 5 [ 
40 143 ' 
From the similar examination of the sodium compound the mole- 
cular weight of albumin was found to be 149G. Other vegetable pro- 
teids examined by Grubler 3 also gave high but variable molecular 
weights. 
Haemoglobin belongs to the proteid compounds capable of crystallisa- 
tion; Zinoffsky 4 prepared haemoglobin crystals from the blood of the 
horse in a very pure state, and the formula calculated for haemoglobin 
from his elementary analyses would be — ■ 
^Ti-jHj^X^O^FeS., 
If a molecule of haematin, C3 2 H 32 N 4 4 Fe, is subtracted, the formula for 
proteid left is — 
*~'&S0^M038-^' 210^2^241 
Jaquet's 5 formula for pure haemoglobin of dog's blood would give the 
proteid molecule a formula — 
( i TT "NT ^ O 
V '7l'C X1 11T1^ > li)4' 3 U 2H 
So that here again there are great discrepancies. 
Such a summary of the principal analyses made, is quite sufficient to 
give point to Drechsel's conclusion, that while divergences of analysis 
exist, even though they are due to extremely small errors, it is futile 
to attempt to measure accurately the size of the proteid molecule. 
Dreehsel points out that in so large a molecule an analytical error of 
- 01 per cent, would have the same importance as one of - l per cent, in 
ordinary analyses. 
It should be added, in conclusion, that some few investigators have 
used the cryoscopic method in attempting the solution of this problem ; 
the molecular weight of egg-albumin by this method is 14,000 
(SabanejefF), 6 of albumoses 1200-2100, and of antipepton much less 
(Paal). 7 
Equally inconclusive, though much more interesting, have been the 
attempts to discover the rational formula for the proteid molecule. The 
that proteids in solution will not pass through a membrane of gelatin or silicic acid, when 
filtered under pressure. The products of proteolysis (proteoses and peptones) will, 
however, pass such a membrane ; the smaller size of their molecules has also been demon- 
strated by the cryoscopic method. Crystalloids pass through such membranes at the same 
rate as water, and can tbus be easily separated from colloids in a solution containing both 
(C. J. Martin, Jour a. Physiol., Cambridge and London, 1896, vol. xx. p. 364). 
1 The subject of vegetable and crystalline proteids will be treated at length in a later 
section of this chapter. 
- Journ. f. prakt. Ohem., Leipzig, 1879, X.F.. Bd. xix. S. 331. 
s Ihid., 1881, Bd. xxiii. S. 97. 
* Ztsckr. f. physiol. Chem., Strassburg, 1885, Bd. x. S. 16. 5 Inaug. Diss., Basel, 1889. 
6 Ber. d. deutsr.h. chem. Gesellsch., Berlin, 1891, Bd. xxiv. Ref. 558. 
7 Ibid. , 1894, Bd. xxvii. S. 1827. For Siegfried's work on the identity of antipeptone 
with a simple compound, which lie has called curate arid, see under " Chemistry of Muscle," 
p. 103. 
